Date of Submission


Date of Award


Institute Name (Publisher)

Indian Statistical Institute

Document Type

Doctoral Thesis

Degree Name

Doctor of Philosophy

Subject Name

Computer Science


Machine Intelligence Unit (MIU-Kolkata)


Kundu, Malay Kumar (MIU-Kolkata; ISI)

Abstract (Summary of the Work)

In image processing and computer vision research, we aim to derive better tools that give us different perspectives on the same image, allowing us to understand not only its content, but also its meaning and significance. Image processing can not compete with the human eye in terms of accuracy but it can outperform the latter easily on observational consistency, and ability to carry out detailed mathematical estimations. With time, image processing research has broadened from the basic pixel-based low- level operations to high-level analysis, that now includes the use of artificially intelligent techniques for image interpretation and understanding. These new technologies are being developed to gain a better semantic understanding of an image based on the relationship between its components and its context.What comprises an image must be first identified before we can analyze the image any further. Texture is a concept used to indicate some spatial properties of image regions. Thus, for an example, we can think of the above adage, where we can identify the printed text portion from the handwritten portion (signature) as two different textures.In the present thesis, we are primarily interested in texture analysis and application of texture analysis to real life data, like document and remotely sensed images and some natural scenes. The recognition of textures in an image is not at all a trivial task. So, this area has been addressed by a number of researchers. Textures have considerable variability in terms of translation, rotation, gray-scale transform etc., and presence of noise, which make their identification difficult.Almost all naturally occurring patterns and natural surfaces exhibit texture. Texture : a fundamental characteristic of an image, and plays an important cue to the human visual system for recognition and interpretation of images. Despite its pivotal role in the analysis of image data, there exists neither a formal/precise definition of texture nor an obvious qualitative measure to characterize it. Image texture can be quanti- tatively expressed in terms of coarseness, fineness, granularity, lineation, randomness and smoothness. The analysis of image texture is extremely important. It requires the understanding of how humans discriminate between different textures and how to model our algorithms to do a predefined task in the best way. Undoubtedly, texture analysis has a wide range of applications. These include:• classifying images based on texture;• segmenting an input image into regions of homogeneous texture;• determining surface-shape on the basis of texture gradient;• synthesizing natural looking textures for graphics applications; andimage retrieval from a database based on texture similarity.Most standard definitions of texture treat it as a measure of coarseness in an image. A simple measure of texture coarseness is based on, first computing the local extremum of an image function along rows and columns. The density of this extremum can then be used as a measure of coarseness of texture. The coarseness at a pixel location can be determined by doing this computation within a small neighborhood.


ProQuest Collection ID:

Control Number


Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.


Included in

Mathematics Commons