Informatics Publishing Limited

Shoba Rani ¹, S. Purushothaman ²

Affiliations
1 Mother Therasa University, Kodaikanal, IN
2 Sun College of Engineering & Technology, Kanyakumari – 629902, Tamil Nadu, IN

Abstract

The work focuses on segmentation of various textures of a given image. In addition, retrieval of the desired component of the image is done. The segmentation and retrieval is based on the texture features of the image. The goal of the work is to develop a promising technique to segment the textures of the image. Basically,we use a set of GABOR filters segmenting the given image and using artificial neural network (ANN) for extracting the relevant information. 

The theme of the work is based on GABOR filters which are based on the famous Gaussian function. Using Gabor filters, does the extractions of features of the various textures in the given image. By using the obtained features, subsequent labeling is done by (Backpropagation algorithm) ANN. The nature of the work involves taking an image texture as input and getting the partitioned or segmented textures as output. It can be very well justified in doing this work as automatic segmentation, identification and classification and as an important role image processing. The future of the work is unlimited.

Keywords

Texture Image Segmentation, Image Retrieval, Gabor Filters, Artificial Neural Network.

Full Text

     

Shoba Rani ¹, S. Purushothaman ²

Affiliations
1 Department of Computer Science, Mother Therasa University, Kodaikanal, IN
2 Sun College of Engineering and Technology, Nagercoil, IN

Abstract

Texture is a ubiquitous experience and can describe a variety of natural phenomena with repetition, such as sound (background noise in a machine room), motion (animal running), visual appearance (surface color and geometry) and human activities (daily lives). Since reproducing the realism of the physical world is a major goal for computer graphics, textures are important for rendering synthetic images and animations. However, as textures are so diverse it is difficult to describe and reproduce them under a common framework. In this paper, new methods for synthesizing textures are presented. The initial part of the paper is concerned with a basic algorithm for reproducing image textures. The limitations of traditional methods can be overcome by the proposed approach based on multi-resolution (search neighborhoods and tree-structured vector quantization) analysis. The paper concerns with various extensions of the basic algorithm; the extensions concentrate on either reproducing textures of different physical phenomena such as motions, or creating textures in novel ways in addition to mimic existing ones.

Keywords

Texture, Wavelet, Multiresolution Analysis and Image Retrieval.

Full Text

     

Shoba Rani ¹, D. Syed Ali ¹

Affiliations
1 CSE, Dr. M.G.R. Educational, and Research Institute University, IN

Abstract

A large amount of single-shot, short videos are created by using personal camcorder in day-to-day life. Many videos are kept in a video pool and merged into a single video. Categorization is done based on transition clues like objects or human beings. For categorization process frame-by-frame search is made on videos in a video pool. Frames are extracted from Video using Viola Jones algorithm. In each frame, complete object is extracted.  Features are extracted from the remaining portion of object using Fisher’s Linear Discriminant function. The features are extracted from objects is considered as a pattern. If 20 frames belonging to a video are considered, then 20 patterns are created. This proposed system is mainly used for separating human beings and objects.

Keywords

Categorization, Integral Image, HAAR Detector, Viola Jones Method, Fisher’s Linear Discriminant Function.

Full Text

     

A. Ambika ¹, K. C. Anu ¹, Shoba Rani ²

Affiliations
1 Dept of CSE, Dr.M.G.R Educational & Research Institute University, IN
2 Dept of CSE, Dr.M.G.R Educational & Research Institute University, IN

Abstract

Liver being the vital organ of human beings, diseases related to liver is considered to be more serious. This paper focuses on various segmentation techniques used to diagnose the abnormality functioning of liver. Image segmentation is considered to be a major process as it involves partitioning the image into smaller segments based on various features that includes intensity value of the image, color, etc.

Keywords

Image Segmentation, Segmentation Techniques, Boundary Tracking, Thresholding.

Full Text

     

Shoba Rani ¹, P. Vijaya Abirami ¹, K. Surya Prakhash ¹

Affiliations
1 Department of Computer Science and Engineering, Dr. M.G.R. Educational and Research Institute University, Chennai – 600095, Tamil Nadu, IN

Abstract

Background/Objective: A computer based biological recognition system implementation for gender identification by using digitally acquired fingerprint through an optical fingerprint sensor. Methods/Statistical Analysis: Gender identification from fingerprint using Discrete Wavelet Transform and Principle Component Analysis for extracting both spatial domain features(SDF) and frequency domain features(FDF), and is combined together for accurate gender identification through a comparison with a set of sample fingerprints of different sex, age, ethnicity, etc. Findings: The results of this approach has revealed that the accuracy of the system depends on the database fingerprint images, bigger the database, better the results, wider the age range of the fingerprints in database, better the results, it also reveals, that unlike the traditional fingerprint ridge to valley thickness ratio methods, the computer based system is more accurate, faster and involves much less human effort, thereby avoiding human errors. Applications/Improvements: This system has a wide scope in anthropology, criminal investigations, personal identification systems like AADHAAR, etc. For large scale applications, a separate algorithm has to be developed for storing the feature vectors of the large amount of database images and significantly improve the processing speed.

Keywords

Discrete Wavelet Transform, Frequency Domain Features (FDF), Principle Component Analysis, Spatial Domain Features (SDF).

Full Text