PROJECT REPORT ON

CONTENT BASED IMAGE RETRIEVAL

BY

1. TEJASVI SHARMA(0909131114) 2. RISHI PATEL(0909131088) 3. RAHUL KUMAR JAISWAL(0909131082)

Under the Guidance of 1. Mrs. Chhaya Grover 2. Mrs. Payal Kaushik

DEPARTMENT OF ELECTRONICS & COMMUNICATION ENGINEERING JSS ACADEMY OF TECHNICAL EDUCATION C-20/1 SECTOR-62, NOIDA April, 2013

Project Report On

CONTENT BASED IMAGE RETRIEVAL

by 1. TEJASVI SHARMA (0909131114) 2. RISHI PATEL ( 0909131088 ) 3. RAHUL KUMAR JAISWAL ( 0909131082 )

Under the Guidance of 1. Mrs.Chhaya Grover 2. Mrs. Payal Kaushik

Submitted to the Department of Electronics & Communication Engineering in partial fulfillment of the requirements for the degree of Bachelor of Technology in Electronics & Communication Engineering JSS Academy of Technical Education, Noida Gautam Buddh Technical University, Lucknow
April 2013

DECLARATION
ii

We hereby declare that this submission is our own work and that, to the best of our knowledge and belief, it contains no material previously published or written by another person nor material which to a substantial extent has been accepted for the award of any other degree or diploma of the university or other institute of higher learning, except where due acknowledgment has been made in the text.

Signature 1. Tejasvi Sharma 0909131114

Signature 2. Rishi Patel 0909131088

Signature 3. Rahul Kumar Jaiswal 0909131082

iii

CERTIFICATE

This is to certify that Project Report entitled Content Based Image Retrieval which is submitted by Tejasvi Sharma, Rishi Patel and Rahul Kumar Jaiswal for partial fulfillment of the requirement for the award of degree B. Tech. in department of Electronics and Communication Engineering of G. B. Technical University, Lucknow is a record of the candidate own work carried out by him under my/our supervision. The matter embodied in this thesis is original and has not been submitted for the award of any other degree.

Signature 1. Mrs. Chhaya Grover Associate Professor

Signature 2. Mrs. Parul Jaiswal Associate Professor

Date:

iv

ACKNOWLEDGEMENT

The successful completion of this report would not have been possible without the help and guidance of many people. We avail this opportunity to convey our gratitude to these people. We would like to show our greatest appreciation to Prof. Dinesh Chandra, Head of Department, Electronics and Communication. We cant say thank you enough for his tremendous support and help. I feel motivated and encouraged every time. Without his encouragement and guidance this report would not have been materialized. We feel privileged to express our deep regards and gratitude to our mentors, Mrs. Chhaya Grover and Mrs. Parul Kaushik without whose active support; this project would not have been possible. We owe sincere regards towards the Department Evaluation Committee , ECE who motivate us to achieve the desired goal.

Signature 1.Tejasvi Sharma 0909131114

Signature 2. Rishi Patel 0909131088

Signature 3. Rahul Kumar Jaiswal 0909131082

ABSTRACT

Content-based image retrieval (CBIR), also known as query by image content (QBIC) and content-based visual information retrieval (CBVIR) is the application of computer vision to the image retrieval problem, that is, the problem of searching for digital images in large databases. "Content-based" means that the search will analyze the actual contents of the image. The term 'content' in this context might refer colors, shapes, textures, or any other information that can be derived from the image itself. Without the ability to examine image content, searches must rely on metadata such as captions or keywords, which may be laborious or expensive to produce. In this project query by example is used which is a query technique that involves providing the CBIR system with an example image that it will then base its search upon. This project is mainly divided into four parts Three search algorithms namely based on colour, shape and texture last part is the application using graphical user interface. Algorithms enlisted in this project are also implemented using Matlab2012b software which is also showing favorable results.

vi

TABLE OF CONTENTS
DECLARATION........................................................................................... CERTIFICATE.............................................................................................. ACKNOWLEDGEMENT............................................................................. ABSTRACT .................................................................................................. LIST OF FIGURES....................................................................................... LIST OF TABLES......................................................................................... LIST OF ABBREVIATIONS. LIST OF SYMBOLS.................................................................................... CHAPTER 1- Introduction 1.1 Background of the problem.......................................................... 1.2 Problem Statement.. CHAPTER 2- Power Supply 2.1. Introduction............................................................................................. 2.2 Bridge Wave Rctifier............................................................................. 2.3 Voltage regulator................................................................................... CHAPTER 3- Color Based Search 3.1 Architecture of CBIR Sysytem ........................ 3.2 Color Based Search Based On Mean And Standard Deviation. 3.2.1 Algorithm 3.2.2 Flowchart for image retrieval based on Mean and S.D... 3.3 Image Retrieval Based on Query in Color Percentage Form 3.3.2 Flow chart image retrieval based on color content. CHAPTER 4- SHAPE BASED IMAGE RETRIEVAL 4.1 Introduction............................................................................................ 4.2. Edge Detection 4.2.1 Sobel Operation 4.2.2 A Prompt Edge Detection Method 4.3. Shape Representation 4.3.1 Central Point Determination 4.3.2 Polar Representation and Distance Sequences 4.3.3 Mountain Climbing Sequence (MCS) vii 20 9 11 12 7 8 8 4 5 4 2 3 Page No. iii iv v vi x xi xii xiii

4.3.4. Shape Matching 4.4 Hausdorff Distance Technique 4.4.1 Introduction 4.4.2. What is Hausdorff distance ? 4.4.3. Application Examples CHAPTER 5- TEXTURE BASED IMAGE RETRIEVAL........... 5.2 Applications 5.2.1. Inspection 5.2.2 Medical Image Analysis 5.2.3. Document Processing 5.3. Texture Analysis 5.3.1. Statistical Methods 5.3.2 Features and offsets used CHAPTER 6- GRAPHICAL USER INTERFACE 6.1 Introduction......................... 6.2 History 6.3 Image Search GUI 41 27 27

5.1 Introduction.............................................................................................

viii

LIST OF FIGURES
Page Figure 1.1:Search Example .... Figure 1.2: Search Example ... Figure 2.1: An Image...... Figure 2.2: Grayscale Image... Figure 2.3: Additive Model of RGB.... Figure 2.4: Additive Model of CMYK. Figure 2.5: Additive Model of Gamut.... Figure 2.6: Connectivity Examples. Figure 3.1: Architecture of CBIR.... Figure 3.2: Algorithm Example.......... Figure 3.3: Example Query.. Figure 3.4: Retrieved Images... Figure 3.5: Color % Example... Figure 4.1: Sobel Operator..... Figure 4.2 : Prompt Edge Detector... Figure 4.3 : Distance and angle of Contour. Figure 4.4: Polar Representation.. Figure 4.5 : Feature Selection.. Figure 4.6,7 : Limitation of min(D(A,B)).... Figure 4.8-14 : Hausdorff Distance Calculation Steps.... Figure 4.15 : Application Example.. Figure 4.16 : Edge Detection... 04 05 06 07 08 09 10 11 13 14 16 16 17 20 20 22 22 23 25 26 28 29

ix

Figure 5.1 : Texture Application.. Figure 5.2 : Texture Analysis.... Figure 5.3 : Stastical Methods. Figure 5.4 : Texture Analysis Example... Figure 5.5 : Texture Features From Power Spectru. Figure 5.6 : Offsets Used.

32 33 38 42 42 43

LIST OF TABLES

Page Table 5.1: Autocorrelation Feature... 16

xi

CHAPTER 1 INTRODUCTION

The use of images for illustration in human communication has its roots millennia ago. Our stone age ancestors painted pictures on the walls of their caves and certainly the use of maps was known in pre-Roman times already. Similar as the increase in printed media through the invention of book-printing, nowadays we observe a rapid increase in digital media caused by an ever increasing growth of computing power, decreasing cost of storage and steady improvement of communication infrastructure technology. But all that information is only useful if one can access it efficiently. This does not only mean fast access from a storage management point of view but also means that one should be able to find the desired information without scanning all information manually. An important part of digital media is image data. In contrast to text, images just consist of pure pixel data with no inherent meaning. Commercial image catalogues therefore use manual annotation and rely on text retrieval techniques for searching particular images. Content-based image retrieval uses the visual contents of an image such as color, shape, texture, and spatial layout to represent and index the image. In typical content-based image retrieval systems , the visual contents of the images in the database are extracted and described by multi-dimensional feature vectors. The feature vectors of the images in the database form a feature database. To retrieve images, users provide the retrieval system with example images or sketched figures. The system then changes these examples into its internal representation of feature vectors. The similarities /distances between the feature vectors of the query example or sketch and those of the images in the database are then calculated and retrieval is performed with the aid of an indexing scheme. The indexing scheme provides an efficient way to search for the image database. Recent retrieval systems have incorporated users' relevance feedback to modify the retrieval process in order to generate perceptually and semantically more meaningful retrieval results. In this we introduce these fundamental techniques for contentbased image retrieval. This project is based on the search performed on three ways, through Colour Shape and Texture. In this part of the report we are dealing only with the colour based search. 1

1.1 Background Of The Problem

Early work on image retrieval can be traced back to the late 1970s. In 1979, a conference on Database Techniques for Pictorial Applications [6] was held in Florence. Since then, the application potential of image database management techniques has attracted the attention of researchers . Early techniques were not generally based on visual features but on the textual annotation of images. In other words, images were first annotated with text and then searched using a text-based approach from traditional database management systems. Comprehensive surveys of early text-based image retrieval methods can be found in. Textbased image retrieval uses traditional database techniques to manage images. Through text descriptions, images can be organized by topical or semantic hierarchies to facilitate easy navigation and browsing based on standard Boolean queries. However, since automatically generating descriptive texts for a wide spectrum of images is not feasible, most text-based image retrieval systems require manual annotation of images. Obviously, annotating images manually is a cumbersome and expensive task for large image databases, and is often subjective, context-sensitive and incomplete. As a result, it is difficult for the traditional text-based methods to support a variety of task-dependent queries. In the early 1990s, as a result of advances in the Internet and new digital image sensor technologies, the volume of digital images produced by scientific, educational, medical, industrial, and other applications available to users increased dramatically. The difficulties faced by text-based retrieval became more and more severe. The efficient management of the rapidly expanding visual information became an urgent problem. This need formed the driving force behind the emergence of content-based image retrieval techniques. In 1992, the National Science Foundation of the United States organized a workshop on visual information management systems to identify new directions in image database management systems. It was widely recognized that a more efficient and intuitive way to represent and index visual information would be based on properties that are inherent in the images themselves. Researchers from the communities of computer vision, database management, human-computer interface, and information retrieval were attracted to this field.

Since then, research on content-based image retrieval has developed rapidly . Since 1997, the number of research publications on the techniques of visual information extraction, organization, indexing, user query and interaction, and database management has increased enormoulsy. Similarly, a large number of academic and commercial retrieval systems have been developed by universities, government organizations, companies, and hospitals. Comprehensive surveys of these techniques and systems can be found .

Content-based image retrieval, uses the visual contents of an image such as color, shape, texture, and spatial layout to represent and index the image. In typical content-based image retrieval systems , the visual contents of the images in the database are extracted and described by multi-dimensional feature vectors. The feature vectors of the images in the database form a feature database. To retrieve images, users provide the retrieval system with example images or sketched figures. The system then changes these examples into its internal representation of feature vectors. The similarities /distances between the feature vectors of the query example or sketch and those of the images in the database are then calculated and retrieval is performed with the aid of an indexing scheme. The indexing scheme provides an efficient way to search for the image database. Recent retrieval systems have incorporated users' relevance feedback to modify the retrieval process in order to generate perceptually and semantically more meaningful retrieval results

1.2 Problem Statement

Commercial image catalogues therefore use manual annotation and rely on text retrieval techniques for searching particular images. However, such an annotation has two main drawbacks: The annotation depends on the person who adds it. Naturally the result may

vary from person to person and furthermore may depend on the context. Within a general image database it may be sufficient to just add an annotation like butterfly whereas this obviously is not sufficient for a biologists database consisting of different types of butterflies only The second problem with manual annotation is that it is very time consuming.

While it may be worthwhile for commercial image collections, it is prohibitive for 3

indexing of images within the World Wide Web. One could not even keep up with the growth of available image data. The image search results, appearing on the first page for text query rose black, are shown in Figure 1

Figure 1.1 Another problem with manual annotation is that of each user may name the query image in their own language which would be hectic to deal with. This method also provides too much responsibility to the end user.It means that it totally depends on human perception of how to name the image and how to form the query text. As perception varies from person to person the method becomes very ineffective.

Figure 1.2 Also there are some images which require greater description which cannot be described by words. Such images may or may not be extracted by this method.

CHAPTER 2 IMAGE PROCESSING

2.1 Introduction
2.1.1 What is an image? An image is an array, or a matrix, of square pixels (picture elements) arranged in columns and rows

. Figure 2.1: An image an array or a matrix of pixels arranged in columns and rows.

In a (8-bit) grayscale image each picture element has an assigned intensity that ranges from 0 to 255. A grey scale image is what people normally call a black and white image, but the name emphasizes that such an image will also include many shades of grey.

Figure 2.2 Each pixel has a value from 0 (black) to 255 (white). 6

The possible range of the pixel values depend on the colour depth of the image, here 8 bit = 256 tones or greyscales. A normal greyscale image has 8 bit colour depth = 256 greyscales. A true colour image has 24 bit colour depth = 8 x 8 x 8 bits = 256 x 256 x 256 colours = ~16 million colours. Some greyscale images have more greyscales, for instance 16 bit = 65536 greyscales. In principle three greyscale images can be combined to form an image with 281,474,976,710,656 greyscales. There are two general groups of images: vector graphics (or line art) and bitmaps (pixel-based or images). Some of the most common file formats are: GIF an 8-bit (256 colour), non-destructively compressed bitmap format. Mostly used for web. Has several sub-standards one of which is the animated GIF. JPEG a very efficient (i.e. much information per byte) destructively compressed 24 bit (16 million colours) bitmap format. Widely used, especially for web and Internet (bandwidthlimited). TIFF the standard 24 bit publication bitmap format. Compresses nondestructively with, for instance, Lempel-Ziv-Welch (LZW) compression. PS Postscript, a standard vector format. Has numerous sub-standards and can be difficult to transport across platforms and operating systems. PSD a dedicated Photoshop format that keeps all the information in an image including all the layers.

2.2 Colours
For science communication, the two main colour spaces are RGB and CMYK. 2.2.1 RGB The RGB colour model relates very closely to the way we perceive colour with the r, g and b receptors in our retinas. RGB uses additive colour mixing and is the basic colour model used in television or any other medium that projects colour with light. It is the basic colour model used in computers and for web graphics, but it cannot be used for print production.The secondary colours of RGB cyan, magenta, and yellow are formed by mixing two of the primary colours (red, green or blue) and excluding the third colour. Red and green combine to make yellow, green and blue to make cyan, and blue and red form magenta. The combination of red, green, and blue in full intensity makes white.In Photoshop using the screen mode for the different layers in an image will make the intensities mix together according to the additive colour mixing model. This is analogous to stacking slide images on top of each other and shining light through them. 7

Figure 2.3: The additive model of RGB. Red, green, and blue are the primary stimuli for human colour perception and are the primary additive colours. 2.2.2 CMYK The 4-colour CMYK model used in printing lays down overlapping layers of varying percentages of transparent cyan (C), magenta (M) and yellow (Y) inks. In addition a layer of black (K) ink can be added. The CMYK model uses the subtractive colour model.

Figure 2.4 8

The colours created by the subtractive model of CMYK don't look exactly like the colours created in the additive model of RGB Most importantly, CMYK cannot reproduce the brightness of RGB colours. In addition, the CMYK gamut is much smaller than the RGB gamut.

2.2.3 Gamut The range, or gamut, of human colour perception is quite large. The two colour spaces discussed here span only a fraction of the colours we can see. Furthermore the two spaces do not have the same gamut, meaning that converting from one colour space to the other may cause problems for colours in the outer regions of the gamut.

Figure 2.5

2.3 Types Of Neighbourhoods

Neighborhood operations play a key role in modern digital image processing. It is therefore important to understand how images can be sampled and how that relates to the various neighborhoods that can be used to process an image. Rectangular sampling In most cases, images are sampled by laying a rectangular grid over an image as illustrated. This results in the type of sampling shown in figure. Hexagonal sampling An alternative sampling scheme is shown in Figure c and is termed hexagonal sampling. 9

Both sampling schemes have been studied extensively and both represent a possible periodic tiling of the continuous image space. We will restrict our attention, however, to only rectangular sampling as it remains, due to hardware and software considerations, the method of choice. Local operations produce an output pixel value b[m=mo ,n=no ] based upon the pixel values in the neighborhood of a[m=mo ,n=no ]. Some of the most common neighborhoods are the 4-connected neighborhood and the 8-connected neighborhood in the case of rectangular sampling and the 6-connected neighborhood in the case of hexagonal sampling illustrated in figure.

Figure 2.6 (a), (b), (c)

10

CHAPTER 3 COLOR BASED SEARCH

Content based Image retrieval is an efficient way to search through an image database by means of image features such as colour, texture, shape, pattern or any combinations of them. Colour is an important cue for image retrieval. The image retrieval based on colour features has proved effective for a large image database, however it is neglected that colour is not a stable parameter, as it depends not only on surface characteristics but also capturing conditions such as illumination, the device characteristics, angle of the device. To retrieve desired images, user has to provide a query image. The system then performs certain feature extraction procedures on it and represents it in the form of feature vectors. The similarities distances between the feature vectors of the query image and those of the images in the database are then calculated and retrieval is performed with the help of indexing schemes. The indexing scheme provides an efficient way to search for the image database. Advances in data storage and image acquisition technologies have enabled the creation of large image datasets. In order to deal with these data, it is necessary to develop appropriate information systems to efficiently manage these collections. Image searching is one of the most important services that need to be supported by such systems. The main reason for color based search is that it is independent of image size and orientation. Also it is one of the most straight-forward features utilized by humans for visual recognition and discrimination. Color is one of the most widely used visual features in content-based image retrieval. It is relatively robust and simple to represent. Various studies of color perception and color spaces have been proposed, in order to find color-based techniques that are more closely aligned with the ways that humans perceive color. The color histogram has been the most commonly used representation technique, statistically describing combined probabilistic properties of the various color channels (such as the Red, Green, and Blue channels), by capturing the number of pixels having particular properties.

11

3.1 Architecture of CBIR Systems

Figure shows a typical architecture of a content-based image retrieval system. Two main functionalities are supported: data insertion and query processing. The data insertion subsystem is responsible for extracting appropriate features from images and storing them into the image database (see dashed modules and arrows). This process is usually performed off-line. The query processing, in turn, is organized as follows: the interface allows a user to specify a query by means of a query pattern and to visualize the retrieved similar images. The queryprocessing module extracts a feature vector from a query pattern and applies a metric (such as the Euclidean distance) to evaluate the similarity between the query image and the database images. Next, it ranks the database images in a decreasing order of similarity to the query image and forwards the most similar images to the interface module. Note that database images are often indexed according to their feature vectors to speed up retrieval and similarity computation. Note that both the data insertion and the query processing functionalities use the feature vector extraction module.

Figure 3.1

12

3.2 Color Based Search Based On Mean And Standard Deviation

The method of image searching and retrieval proposed here mainly focuses on the generation of the feature vector by calculating the average means. Steps of the algorithm are given below. 3.2.1 Algorithm Step1: Three color planes namely Red , Green and Blue are separated. Step2: For each plane row mean and column mean of colours are calculated. Pictorially the average row and column mean is calculated as follows The average mean is calculated by adding up the mean of every row. E.g.: (3+8+13+18+23) = 13. Similarly with column it would be (11+12+13+14+15) = 13 shown as in fig given below.

Figure 3.2 Step3: The average of all row means and all columns means is calculated for each plane. Step4: The features of all 3 planes are combined to form a feature vector. Once the feature vectors are generated for all images in the database, they are stored in a feature database Step5: The Euclidian distances between the feature vector of query image and the feature database are calculated using Eq given below.

13

The value of d is calculated by summation of the squares of difference of the features of database image and query image as mentioned in figure. 3.2.2 Flowchart for image retrieval based on Mean and S.D. Start Select the query image Convert query and database images from RGB to Gray Calculate mean and standard deviation of the query and database images Compare the database and query images on the basis of mean and sd Retrieve the images into a separate directory

Stop To discuss the performance of the algorithm we have considered the beach class as an example. The query image for the same is shown in the Fig This query image is used for both the databases. Query Image The algorithm is applied on the first database of 300 images to generate feature vector for each image in the database and the query image and hence calculate the Euclidian distance to find the better match. The algorithm has produced very good results as it can be seen in the fig below where the first 20 retrieved images found are shown.

14

Figure 3.3 Query

Figure 3.4 Retrieved Images

Retrieval Efficiency The common evaluation measures used in CBIR systems are precision, defined as Precision =No of relevant images retrieved/Total number of images retrieved Recall =No.of relevant images retrieved/Total no.of relevant images in the database.

15

3.3 Image Retrieval Based on Query in Color Percentage Form

To make the algorithm clear we will first start with the example as shown below

Figure 3.5

3.3.1 Algorithm: Step 1: Get the input for the required colour content according to which the image from our database is to be sorted out. Step 2: Extract the RGB metrics of each images in the database Step 3: Compare the RGB content extracted above with the query colour content Step 4: Extract those images which nearly satisfy whose colour content nearly satisfy with the input colour content. Step 5: Retrieve the images into separate direct

16

3.3.2 Flow chart image retrieval based on color content Start Set the required color content Extract RGB intensity matrics from the data base images

Compare the required content with that of database images

Retrieve the images into a separate directory

Stop

17

CHAPTER 4 SHAPE BASED IMAGE RETRIEVAL

4.1 Introduction
Because of recent advances in computer technology and the revolution in the way information is processed, increasing interest has developed in automatic information retrieval from huge databases. In particular, content-based retrieval has received considerable attention and, consequently, improving the technology for content-based querying systems becomes more challenging. There are four important feature components for content-based image retrieval: color, texture, shape. One of the most challenging aspects to content based image retrieval is retrieval by shape. No mathematically rigorous definition of shape similarity accounts for the various semantic qualities that humans assign to shapes.1 Without textual tags or other semantic descriptors, we can only approximate shape similarity from the 2D image. There have been a variety of proposed methods for determining similarity between planar shapes, including moment-based matching algorithms hausdorff-distance based metrics,4,5 and schemes for matching turning angle around the perimeter of a shape.6 The goal of any of these algorithms is to approximate the perceptual judgments of the user. Algorithms which more closely approximate the perceptual judgments of the user are considered to be more successful.

4.2. Edge Detection

Edge detection is a necessary pre-processing step in most of computer vision and image understanding systems. The accuracy and reliability of edge detection is critical to the overall performance of these systems. Earlier researchers paid a lot of attention to edge detection, but up to now, edge detection is still highly challenging. In this section, we will briefly illustrate two common edge detection methods, and point out their drawbacks. In addition, we introduce a simple and efficient method for edge detection.

4.2.1 Sobel Operation Another common edge detection method is the sobel method. The sobel operation uses four masks to detect edge points. The sobel operation computes the partial derivatives of a local pixel to form the gradient. For the gray-level function f(x,y), the gradient of f at pixel 18

(x,y) is defined as a vector. The magnitude of f can be approximated as the following equation:

Figure 4.1 4.2.2 A Prompt Edge Detection Method In this subsection, an easy and effective edge detection method, the Prompt edge detection method, is introduced. The Prompt edge detector detects edge points by checking the differences in the gray value for each point from those of its neighbors. For a pixel (x,y) with gray value g(x,y), let g0(x,y), g1(x,y), , and g7(x,y) denote the gray values of its neighbors in 8 directions .

Figure 4.2 Let hd (x, y) = g(x, y) - gd (x, y) be the difference in the gray value of pixel (x,y) from that of its neighbour in direction d. Let B d(x,y) be the number of differences that exceed a threshold T, where T = a + c, c is a constant, and a is the average difference between all pairs of adjacent pixels' gray values in the image. In this work, we take the value 2 for c and, instead of taking a single value for T, we take the local average differences for T by dividing the hole image into M M blocks and computing the average difference for each block. The pixel (x,y) is indicated as an edge point if the following inequalities hold : 3 Bd (x, y) 6 These inequalities can also avoid indicating noisy points as edge points.

4.3. Shape Representation

A good representation can manifest an objects characteristics. Furthermore, it would help achieve a high recognition rate for a content-based image retrieval system. Some researchers have shown that objects are very distinguishable based on their visible features. Among these features, shape is the most important for recognizing objects. In this section, we 19

introduce a new shape representation method that is invariant to translation, rotation, and scaling, and also has high tolerance to complex, occluding, or even deformed images.

4.3.1 Central Point Determination The first step in shape representation for an object is to locate the central point of the object. To permit invariance to translation, rotation and scaling, the geometric center of the object is selected as a reference point. We use the equation below to compute the centroid of an object.

where n is the number of points of an object.

4.3.2 Polar Representation and Distance Sequences A shape representation method is outlined in this section. We characterize the contour using a sequence of contour points described in polar from. Here, we take the pole at the centroid (xc,yc), then the contour graph can be represented by a polar equation d = f ( ) , and each contour point (x,y) has the polar description (d, ) , where x, y, d, and are related using Equations given below

A sketch map is shown of Distance and angle of the contour points relative to centroid

Figure 4.3

20

With the polar description, we may represent contour points as a sequence (d0 , d1,, dn1) , where di = f ( i) i =i 2 /n , where di is known to be the distance between the

corresponding contour point and the centroid. Figure 3.4 illustrates the graph of the polar equation d = f ( ) for a given contour.

Figure 4.4 An example of polar representation (a) A contour of an object. (b) Graph of polar representation for the contour in (a). We can obtain the distance sequence ( d0 , d1,, dn-1) by successively rotating a ray emanating from the pole counter clockwise through an fixed angle , where = 2 n for a positive integer n, and at each step of rotation. Recording the distance of intersection point of the current ray lies with the contour . This method of representing and describing the shape of an object is adequate for convex objects. However, for an object containing concavities, a ray may intersect the contour with more than one point. If we only record the distance of one point (say, smallest distance) at each step of rotation, the contour points of some protrusions may be missed. We can eliminate this problem by recording the distance of the farthest intersection point at each rotating the scanning ray step, or for a more detailed description, recording the distances, or furthermore associating the number of all the intersection points. To provide scale invariance, the maximum distance is computed and all distances are normalized to it. Thus, all values fall between 0 and 1 regardless of how much the objects are scaled.

Figure 4.5 Features selection of two different concave objects. 21

4.3.3 Mountain Climbing Sequence (MCS) Since we used the pole at the centroid of the object, the representation of a distance sequence is translation invariant. To achieve rotation-invariance, Bernier rotated the polar coordinate system, such that the maximum distance was associated with the angle of zero. However there is no any guarantee for this method if there are more than one maximum distance. The representation will not be unique. In the example shown in figure, both objects have three maximal distances and each has three possible representations: ( 2 , 1, 1, 2 , 1, 2 ,1), ( 2 , 1, 2 ,1, 2 , 1,1), and ( 2 ,1 2 , 1,1, 2 ,1). To reduce this problem, we propose another representation method that deals with all distances rather than only the individual maximal distances in the sequence. First, we evaluate the ordering-consistency functions ci's, as defined in Equation below, at the sequence D = (d0, d1, , dn-1) as described , and determine the index, s, of the function having the smallest value, as defined in Equation ci. The distances in the sequence D are then shifted forward for s positions to yield a new sequence Dm= (ds,ds+1, ,ds+n-1), where dk = dk mod n for each k. This new sequence Dm is called the mountain climbing sequence(MCS) for the object. Sequence (MCS) for the object. The MCS representations for Figures 3.5(a) and (b) are identical and both come out as (1, 1, 2 , 1, 2 , 1, 2 ), and as a result, they are considered having the same shape.

3.3.4. Shape Matching

The final step of content-based image retrieval is the matching stage. As discussed in the receding subsection, we use the MCS to represent the shape of objects. We proceed to the matching stage to measure the similarity between two objects by simply evaluating the Euclidean distance between each objects MCSs of them. The Euclidean distance between two sequences is given in Equation below

22

where U and V are the query and dababase images, vi, ui are their ith features, respectively, and n is the dimension of the feature space. Any arbitrary two dimensional shape can then be compared to any other using the outline method.

4.4 Hausdorff Distance Technique

4.4.1 Introduction Although hausdorff distance technique is very robust nature application of which resulted in the appearance of time delay. So in order to get better result we have to reach a trade-off between time delay accuracy.

That is for every point a of A, find its smallest distance to any point of B; finally keep the smallest distance found among point a. There is a limitation for the above calculated distance given as below:

Figure 4.6 First, shortest distance doesnt consider the whole shape

23

Figure 4.7 Second, shortest distance doesnt account for the position of the object

4.4.2. What is Hausdorff distance ? Named after Felix Hausdorff (1868-1942), Hausdorff distance is the maximum distance of a set to the nearest point in the other set . More formally, Hausdorff distance from set A to set B is a maxi-min function, defined as

where a and b are points of sets A and B respectively, and d(a, b) is any metric between these points ; for simplicity, we'll take d(a, b) as the Euclidian distance between a and b. If for instance A and B are two sets of points, steps to to understand how hausdorff distance is calculated are shown below :

Step 1:

Figure 4.8

24

Given two sets of points A and B ,finding h(A,B)

Step 2:

Figure 4.9 Compute distance between ai and bjs. Step3:

Figure 4.10

Keeping the shortest distance.

Step4:

Figure 4.11 Compute distance between a2 and bjs. 25

Step5:

. Figure 4.12 And keep the shortest distance

Step6:

Figure 4.13 Finally , find the largest of the two distances. Step 7:

Figure 4.14

This is h(A,B). 26

4.4.3. Application Examples One of the main application of the Hausdorff distance is image matching, used for instance in image analysis, visual navigation of robots, computer-assisted surgery, etc. Basically, the Hausdorff metric will serve to check if a template image is present in a test image ; the lower the distance value, the best the match. That method gives interesting results, even in presence of noise or occlusion (when the target is partially hidden).

Say the small image below is our template, and the large one is the test image :

Figure 4.15

We want to find if the small image is present, and where, in the large image. The first step is to extract the edges of both images, so to work with binary sets of points, lines or polygons :

27

Figure 4.16

Edge extraction is usually done with one of the many edge detectors known in image processing, such as Canny edge detector, Laplacian, Sobel, etc. minimum Hausdorff distance between two images found a best match .

28

CHAPTER 5 TEXTURE BASED IMAGE RETRIEVAL

5.1 Introduction
In many machine vision and image processing algorithms, simplifying assumptions are made about the uniformity of intensities in local image regions. However, images of real objects often do not exhibit regions of uniform intensities. For example, the image of a wooden surface is not uniform but contains variations of intensities which form certain repeated patterns called visual texture. The patterns can be the result of physical surface properties such as roughness or oriented strands which often have a tactile quality, or they could be the result of reectance differences such as the color on a surface. We recognize texture when we see it but it is very difcult to dene. This difculty is demonstrated by the number of different texture denitions attempted by vision researchers. Some have compiled a catalogue of texture denitions in the computer vision literature and we give some examples here. We may regard texture as what constitutes a macroscopic region. Its structure is simply attributed to the repetitive patterns in which elements or primitives are arranged according to a placement rule. A region in an image has a constant texture if a set of local statistics or other local properties of the picture function are constant, slowly varying, or approximately periodic. An image texture is described by the number and types of its (tonal) primitives and the spatial organization or layout of its (tonal) primitives. A fundamental characteristic of texture: it cannot be analyzed without a frame of reference of tonal primitive being stated or implied. For any smooth gray-tone surface, there exists a scale such that when the surface is examined, it has no texture. Then as resolution increases, it takes on a fine texture and then a coarse texture. The notion of texture appears to depend upon three ingredients: (i) some local order is repeated over a region which is large in comparison to the orders size, (ii) the order consists in the arrangement of elementary parts, and (iii) the parts are roughly uniform entities having approximately the same dimensions everywhere within the textured region.

29

This collection of denitions demonstrates that the denition of texture is formulated by different people depending upon the particular application and that there is no generally agreed upon denition. Some are perceptually motivated, and others are driven completely by the application in which the denition will be used. Image texture, dened as a function of the spatial variation in pixel intensities (gray values), is useful in a variety of applications and has been a subject of intense study by many researchers. One immediate application of image texture is the recognition of image regions using texture properties.

5.2 Applications Texture analysis is an important and useful area of study in machine vision. Most natural surfaces exhibit texture and a successful vision system must be able to deal with the textured world surrounding it. This section will review the importance of texture perception from two viewpoints from the viewpoint of human vision or psychophysics and from the viewpoint of practical machine vision applications. Texture analysis methods have been utilized in a variety of application domains. In some of the mature domains (such as remote sensing) texture already has played a major role, while in other disciplines (such as surface inspection) new applications of texture are being found. We will briey review the role of texture in automated inspection, medical image processing, document processing, and remote sensing. Images from two application domains are shown in Figure 1. The role that texture plays in these examples varies depending upon the application. For example, in the SAR images of Figures 4.1(b) and (c) texture is dened to be the local scene heterogeneity and this property is used for classication of land use categories such as water, agricultural areas, etc. In the ultrasound image of the heart in Figure 4.1(a), texture is dened as the amount of randomness which has a lower value in the vicinity of the border between the heart cavity and the inner wall than in the blood lled cavity. This fact can be used to perform segmentation and boundary detection using texture analysis methods.

30

Figure 5.1

5.2.1. Inspection There has been a limited number of applications of texture processing to automated inspection problems. These applications include defect detection in images of textiles and automated inspection of carpet wear and automobile paints. Texture pairs with identical second-order statistics. The bottom halves of the images consist of texture tokens that are different from the ones in the top half. (a) Humans cannot perceive the two regions without careful scrutiny. (b) The two different regions are immediately discriminable by humans. In the detection of defects in texture images, most applications have been in the domain of textile inspection. Dewaele used signal processing methods to detect point defects and line defects in texture images. They have sparse convolution masks in which the bank of lters are adaptively selected depending upon the image to be analyzed. Texture features are computed from the ltered images. A Mahalanobis distance classier is used to classify the defective areas. Some have dened a simple window differencing operator to the texture features obtained from simple ltering operations. This allows one to detect the boundaries of defects in the texture. Chen and Jain used a structural approach to defect detection in textured images. They extract a skeletal structure from images, and by detecting anomalies in certain statistical features in these skeletons, defects in the texture are identied. Conners utilized texture analysis

methods to detect defects in lumber wood automatically. The defect detection is performed by 31

dividing the image into sub windows and classifying each sub window into one of the defect categories such as knot, decay, mineral streak, etc. The features they use to perform this classication is based on tonal features such as mean, variance, skewness, and kurtosis of gray levels along with texture features computed from gray level co-occurrence matrices in analyzing pictures of wood. The combination of using tonal features along with textural features improves the correct classication rates over using either type of feature alone. In the area of quality control of textured images, Siew proposed a method for the assessment of carpet wear. They used simple texture features that are computed from second-order gray level dependency statistics and from rst-order gray level difference statistics. They showed that the numerical texture features obtained from these techniques can characterize the carpet wear successfully. Jain used the texture features computed from a bank of Gabor lt ers to automatically classify the quality of painted metallic surfaces. A pair of automotive paint nish images is shown where the image in (a) has uniform coating of paint, but the image in (b) has mottle or blotchy appearance.

Figure 5.2 Examples of images from various application domains in which texture analysis is important.

5.2 Medical Image Analysis Image analysis techniques have played an important role in several medical applications. In general, the applications involve the automatic extraction of features from the image which are then used for a variety of classication tasks, such as distinguishing normal tissue from abnormal tissue. Depending upon the particular classication task, the extracted features capture morphological properties, color properties, or certain textural properties of the image.

32

The textural properties computed are closely related to the application domain to be used. For example, Sutton and Hall discuss the classication of pulmonary disease using texture features. Some diseases, such as interstitial brosis, affect the lungs in such a manner that the resulting changes in the X-ray images are texture changes as opposed to clearly delineated lesions. In such applications, texture analysis methods are ideally suited for these images. Sutton and Hall propose the use of three types of texture features to distinguish normal lungs from diseased lungs. These features are computed based on an isotropic contrast measure, a directional contrast measure, and a Fourier domain energy sampling. In their classication experiments, the best classication results were obtained using the directional contrast measure. Harms used image texture in combination with color features to diagnose leukemic malignancy in samples of stained blood cells. They extracted texture micro-edges and textons between these micro-edges. The textons were regions with almost uniform color. They extracted a number of texture features from the textons including the total number of pixels in the textons which have a specic color, the mean texton radius and texton size for each color and various texton shape features. In combination with color, the texture features signicantly improved the correct classication rate of blood cell types compared to using only color features. Landeweerd and Gelsema extracted various rst-order statistics (such as mean gray level in a region) as well as second-order statistics (such as gray level cooccurrence matrices) to differentiate different types of white blood cells. Insana used textural features in ultrasound images to estimate tissue scattering parameters. They made signicant use of the knowledge about the physics of the ultrasound imaging process and tissue characteristics to design the texture model. Chen used fractal texture features to classify ultrasound images of livers, and used the fractal texture features to do edge enhancement in chest X-rays. Lundervold used fractal texture features in combination with other features (such as response to edge detector operators) to analyze ultrasound images of the heart (see Figure 5.1). The ultrasound images in this study are time sequence images of the left ventricle of the heart. Figure 5.1 shows one frame in such a sequence. Texture is represented as an index at each pixel, being the local fractal dimension within an window estimated according to the fractal Brownian motion model proposed by Chen. The texture feature is used in addition to a number of other traditional features, including the response to a Kirsch edge operator, the gray level, and the result of temporal operations. The fractal dimension is expected to be higher on an average in blood than in tissue due to the noise and backscatter characteristics of 33

the blood which is more disordered than that of solid tissue. In addition, the fractal dimension is low at non-random blood/tissue interfaces representing edge information.

5.2.3. Document Processing One of the useful applications of machine vision and image analysis has been in the area of document image analysis and character recognition. Document processing has applications ranging from postal address recognition to analysis and interpretation of maps. In many postal document processing applications (such as the recognition of destination address and zip code information on envelopes), the rst step is the ability to separate the regions in the image which contain useful information from the background. Most image analysis methods proposed to date for document processing are based upon the characteristics of printed documents and try to take advantage of these properties. For example, generally newspaper print is organized in rectangular blocks and this fact is used. The processing of the ultrasound images of the heart using textural features.. Many methods work on images based on precise algorithms which one might consider as having morphological characteristics. For example, Wang and Srihari used projection proles of the pixel values to identify large text blocks by detecting valleys in these proles. Wahl used constrained run lengths and connected component analysis to detect blocks of text. Fletcher and Kasturi used the fact that most text blocks lie in a straight line, and utilized Hough transform techniques to detect collinear elements. Taxt view the identication of print in document images as a two-category classication problem, where the categories are print and background. They use various classication methods to compute the segmentation including contextual classication and relaxation algorithms. One can also use texture segmentation methods for preprocessing document images to identify regions of interest. An example of this can be seen in Figure. The texture segmentation algorithm described in was used to segment a newspaper image. In the resulting segmentation, one of the regions identied as having a uniform texture, which is different from its surrounding texture, is the bar code block. A similar method is used for locating text blocks in newspapers. A segmentation of the document image is obtained using three classes of textures: one class for the text regions, a second class for the uniform regions that form the background or images where intensities vary slowly, and a third class for the transition areas between the two types of regions.

34

5.3. Texture Analysis

Approaches to texture analysis are usually categorised into structural, statistical, model-based and transform methods.

Structural approaches represent texture by well defined primitives (micro texture) and a hierarchy of spatial arrangements (macro texture) of those primitives. To describe the texture, one must define the primitives and the placement rules. The choice of a primitive (from a set of primitives) and the probability of the chosen primitive to be placed at a particular location can be a function of location or the primitives near the location. The advantage of the structural approach is that it provides a good symbolic description of the image; however, this feature is more useful for synthesis than analysis tasks. The abstract descriptions can be ill defined for natural textures because of the variability of both micro- and macrostructure and no clear distinction between them. A powerful tool for structural texture analysis is provided by mathematical morphology . It may prove to be useful for bone image analysis, e.g. for the detection of changes in bone microstructure. In contrast to structural methods, statistical approaches do not attempt to understand explicitly the hierarchical structure of the texture. Instead, they represent the texture indirectly by the non-deterministic properties that govern the distributions and relationships between the grey levels of an image. Methods based on second-order statistics (i.e. statistics given by pairs of pixels) have been shown to achieve higher discrimination rates than the power spectrum (transform-based) and structural methods. Human texture discrimination in terms of texture statistical properties is investigated Julesz. Accordingly, the textures in grey-level images are discriminated spontaneously only if they differ in second order moments. Equal second order moments, but different third-order moments require deliberate cognitive effort. This may be an indication that also for automatic processing, statistics up to the second order may be most important. The most popular second-order statistical features for texture analysis are derived from the so-called co-occurrence matrix. They were demonstrated to feature a potential for effective texture discrimination in biomedical-images. 35 The approach based on

multidimensional co-occurrence matrices was recently shown to outperform wavelet packets (a transform-based technique) when applied to texture classification. Model based texture analysis using fractal and stochastic models, attempt to interpret an image texture by use of, respectively, generative image model and stochastic model. The parameters of the model are estimated and then used for image analysis. In practice, the computational complexity arising in the estimation of stochastic model parameters is the primary problem. The fractal model has been shown to be useful for modelling some natural textures. It can be used also for texture analysis and- 3 - discrimination; however, it lacks orientation selectivity and is not suitable for describing local image structures. Transform methods of texture analysis, such as Fourier, Gabor represent an image in a space whose co-ordinate system has an interpretation that is closely related to the characteristics of a texture (such as frequency or size). Methods based on the Fourier transform perform poorly in practice, due to its lack of spatial localisation. Gabor filters provide means for better spatial localisation; however, their usefulness is limited in practice because there is usually no single filter resolution at which one can localise a spatial structure in natural textures. Compared with the Gabor transform, the wavelet transforms feature several advantages. Identifying the perceived qualities of texture in an image is an important rst step towards building mathematical models for texture. The intensity variations in an image which characterize texture are generally due to some underlying physical variation in the scene (such as pebbles on a beach or waves in water). Modelling this physical variation is very difcult, so texture is usually characterized by the two-dimensional variations in the intensities present in the image. This explains the fact that no precise, general denition of texture exists in the computer vision literature. In spite of this, there are a number of intuitive properties of texture which are generally assumed to be true.

5.3.1. Statistical Methods One of the dening qualities of texture is the spatial distribution of gray values. The use of statistical features is therefore one of the early methods proposed in the machine vision literature. In the following, we will use to denote an N X N

image with G gray levels. A large number of texture features have been proposed. But, these features are not independent as pointed out by Tomita. The relationship between the various statistical texture measures and the input image is summarized in Fig.4.3. Picard has also related the gray level co-occurrence matrices to the Markov random eld models. 36

Figure 5.3 Co-occurrence Matrices Spatial gray level co-occurrence estimates image properties related to second-order statistics. Haralick suggested the use of gray level co-occurrence matrices (GLCM) which have become one of the most well-known and widely used texture features. The G X G gray level cooccurrence matrix Pd for a displacement vector d= (dx,dy) is dened as follows. The entry (i,j) of Pd is the number of occurrences of the pair of gray levels i and j which are distance d apart. Formally given as, Pd( I, j) =|{((r, s),(t, v):I(r, s)=I, I(t,v)=j}| ` and |.| is the cardinality of a set.

As an example, consider the following image containing 3 different gray values:

1100 1100 0022 0022

37

The gray level co-occurrence matrix for this image for a displacement vector d= (1,0) is given as follows: 402 Pd= 220 002

Here the entry (0,0) of Pd is 4 because there are four pixel pairs (0,0) that are offset by amount. Examples of Pd for other displacement vectors is given below.

Notice that the co-occurrence matrix so dened is not symmetric. But a symmetric cooccurrence matrix can be computed by the formula P = Pd+P-d . The co-occurrence matrix reveals certain properties about the spatial distribution of the gray levels in the texture image. For example, if most of the entries in the co-occurrence matrix are concentrated along the diagonals, then the texture is coarse with respect to the displacement vector d. Haralick has proposed a number of useful texture features that can be computed from the co-occurrence matrix. Table 1 lists some of these features.

The co-occurrence matrix features suffer from a number of difculties. There is no well established method of selecting the displacement vector and computing co-occurrence matrices for different values of is not feasible. For a given , a large number of features can be computed from the co-occurrence matrix. This means that some sort of feature selection method must be used to select the most relevant features. The co-occurrence matrix-based texture features have also been primarily used in texture classication tasks and not in segmentation tasks Autocorrelation Features 38

An important property of many textures is the repetitive nature of the placement of texture elements in the image. The autocorrelation function of an image can be used to assess the amount of regularity as well as the neness/coarseness of the texture present in the image. Formally, the autocorrelation function of an image is dened as follows:

The image boundaries must be handled with special care but we omit the details here. This function is related to the size of the texture primitive (i.e., the neness of the texture). If the texture is coarse, then the autocorrelation function will drop off slowly; otherwise, it will drop off very rapidly. For regular textures, the autocorrelation function will exhibit peaks and valleys. The autocorrelation function is also related to the power spectrum of the Fourier transform (see Figure). Consider the image function in the spatial domain I(x,y) and its Fourier Transform F(u,v) . The quantity |F(u,v)|2 is dened as the power spectrum where|.| is the modulus of a complex number. The example in Figure 4.4 illustrates the effect of the directionality of a texture on the distribution of energy in the power spectrum. Early

Table 5.1

39

Figure 5.4

Figure 5.5 Texture features computed from the power spectrum of the image approaches using such spectral features would divide the frequency domain into rings (for frequency content) and wedges (for orientation content) as shown in Figure 4.5. The

40

frequency domain is thus divided into regions and the total energy in each of these regions is computed as texture features. 5.3.2 Features and offsets used After you create the GLCMs, you can derive several statistics from them . These statistics provide information about the texture of an image. The following table lists the statistics.

Offsets used are shown in the next figure 5.6

Figure 5.6

41

CHAPTER 6 GRAPHICAL USER INTERFACE

6.1 Introduction
A graphical user interface (GUI) is a human-computer interface (i.e., a way for humans to interact with computers) that uses windows, icons and menus and which can be manipulated by a mouse and often to a limited extent by a keyboard as well. A window is a rectangular portion of the monitor screen that can display its contents (e.g., a program, icons, a text file or an image) seemingly independently of the rest of the display screen. A major feature is the ability for multiple windows to be open simultaneously. Each window can display a different application, or each can display different files (e.g., text, image or spreadsheet files) that have been opened or created with a single application. An icon is a small picture or symbol in a GUI that represents a program , a file, a directory or a device (such as a hard disk or floppy). Icons are used both on the desktop and within application programs. Examples include small rectangles (to represent files), file folders (to represent directories), a trash can (to indicate a place to dispose of unwanted files and directories) and buttons on web browsers (for navigating to previous pages, for reloading the current page, etc.). Commands are issued in the GUI by using a mouse, trackball or touchpad to first move a pointer on the screen to, or on top of, the icon, menu item or window of interest in order to select that object. Then, for example, icons and windows can be moved by dragging (moving the mouse with the held down) and objects or programs can be opened by clicking on their icons. Advantages of GUIs A major advantage of GUIs is that they make computer operation more intuitive, and thus easier to learn and use. For example, it is much easier for a new user to move a file from one directory to another by dragging its icon with the mouse than by having to remember and type seemingly arcane commands to accomplish the same task. Adding to this intuitiveness of operation is the fact that GUIs generally provide users with immediate, visual feedback about the effect of each action. For example, when a user deletes an icon representing a file, the icon immediately disappears, confirming that the file has been deleted (or at least sent to the trash can). This contrasts with the situation for a CLI, in which the user types a delete command (inclusive of the name of the file to be deleted) but receives no automatic feedback indicating that the file has actually been removed. In addition, GUIs allow users to take full advantage of the powerful multitasking (the ability for multiple programs and/or multiple instances of single programs to run simultaneously) capabilities of modern operating systems by allowing such multiple programs and/or instances to be displayed simultaneously. The result is a large increase in the flexibility of computer use and a consequent rise in user productivity. 42

But the GUI has become much more than a mere convenience. It has also become the standard in human-computer interaction, and it has influenced the work of a generation of computer users. Moreover, it has led to the development of new types of applications and entire new industries. An example is desktop publishing, which has revolutionized (and partly wiped out) the traditional printing and typesetting industry.

6.2 History
The origin of the GUI can be traced back to Vannevar Bush, a scientist and futurist who worked at the Massachusetts Institute of Technology (MIT) during World War II. In his now classic 1945 magazine article As We May Think, Bush proposed an information administration tool, the Memex , that would allow data to be stored on microfilm and made easily accessible, linkable with hyperlinks and programmable. In 1963 Ivan Sutherland, a graduate student at MIT, developed a program for his Ph.D. dissertation called Sketchpad, which allowed the direct manipulation of graphic objects on a CRT screen using a light pen. His concept included the capability to zoom in and out on the display, the provision of memory for storing objects and the ability to draw precision lines and corners on the screen. Much additional progress occurred at Xerox Palo Alto Research Center (PARC), which Xerox Corporation established in 1970 in Palo Alto, California for the purpose of creating "the architecture of information" and "humanizing computers." This included developing the first usable GUI, which was incorporated into PARC's Alto computer. Steve Jobs, the co-founder of Apple Computer, was heavily inspired by the innovations at nearby PARC, and he thus decided to incorporate a GUI into his company's computers. Apple considerably extended PARC's work, including developing overlapping windows, manipulable icons, a fixed menu bar, dropdown menus and a trash can. The Apple Macintosh, launched in 1984, was the first commercially successful use of a GUI. It was so successful, in fact, that the GUI was subsequently adopted by most other developers of operating systems and application software, and it is now used on (or at least available for) virtually all types of computers. Microsoft announced development of its first operating system that incorporated a GUI in November 1983, and the initial version, Windows 1.0, was released in November 1985. Windows 2.0, released in December 1987, represented a major improvement over the primitive Windows 1.0 with its addition of icons and overlapping windows, but it was not until 1995 with the launching of Windows 95 that Microsoft was able to offer a relatively high quality GUI. The Future One of the most interesting areas of exploration is GUIs that provide the user with the illusion of navigating through a three-dimensional space. Another area of research is increasing user control over GUI objects, such as being able to rotate and freely change the size and transparency of icons. Also being studied is the increased use of natural languages to interact with computers as a supplement or complement to GUIs. 43

6.3 Image Search GUI

The GUI of this project deals with the formation of a user friendly window that allows him to effectively search the database and extract the required images out of it. For this pupose it has several keys and buttons which are to be discussed next.

ALGORITHM 1-GUI Step 1- The user has to give the complete URL of the query image. Step-2 The program then requires the range of averages and standard deviations that are tolerable for the user. This can be input through the edit text windows provided in GUI. Step-3 The user can then press the button search for starting the database search. Step-4 User has another option of whether to display the images or store them in a particular folder location. ALGORITHM 2-GUI Step 1- The user has to provide the percentage of different color intansities in the required image as query. Step-2 The program requires the percentages in the edit boxes created especially for this purposes.eg-Provide red intensity in R box. Step-3 The user can then press the button search for starting the database search. Step-4 User has another option of whether to display the images or store them in a particular folder location.

44