Digital Image Processing

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