Projects

Some details on some of the things I've worked on in the recent past: (in reverse chronological order)

Face Detection & Recognition
The objective of this effort is to detect a face in an image and to recognize it. Images containing faces are essential to intelligent vision-based human computer interaction systems, and research efforts in face processing include face recognition, face tracking, pose estimation and expression and gender recognition. The problem of detecting and recognizing human faces under unsupervised environments is under study. The idea is to extract all the information that correspond to a face and utilize these as the parameters for face recognition. This is going to be one of the "most challenging" projects I have ever attempted. A Knowledge-Based Visual DBMS
When multimedia databases are the norm, traditional methods of working with databases no longer apply. The database management systems need to reassess existing data handling paradigms to accommodate these new forms of non-textual data, as well as new types of operations that processing such data requires. This project tries to explain the architecture involved in the development of a knowledge based database system; where the inherent knowledge (cognitive aspects) from the images is used to query the database by using suitable image processing algorithms instead of feeding in textual descriptions associated with each entity. A particular example of facial database system has been taken as an example to illustrate the extraction of knowledge from the images that contain faces and suitably presented at the time of search and retrieval.

A Proposed “Bee-Hive” Structure for Digital Images
This paper is concerned with a change in the representation of the structure of the images where an efficient "Bee-Hive" structure has been proposed. This will require a complete revolutionary change in the hardware representation, interconnectivity and performance with a corresponding increase in the level of parallelism, since at an instant more number of pixels can be considered and thus the degree of parallelism is increased. The new system of representation has numerous advantages over the existing systems. Edge Detection Techniques
Edge detection, is based upon the detection of local discontinuities which mainly correspond to the boundaries of objects in the image. A review of various edge detection techniques for object detection and its attributes verification is undergoing at present. The problem with visual perception and gradient detection is under study.This is mainly concerned with ROBOTIC VISION.

Review of Quantum Computing
The objective of this effort was to accelerate the understanding of the science of quantum computing and the development and demonstration of quantum devices, and algorithms needed for "a radically new computer architecture". Apart from mentioning the principles of quantum computing ,an overview of the way in which quantum computing will speed-up processing, the key areas in which this technology will be applied, it's advantages and shortcomings was also studied. Steg-Analysis of LSB affected Images
The purpose of this effort was to analyze the nature of LSB affected images, corrupted due to steganography. A passive warden approach to the Prisoner Problem was initiated. The idea was to not develop a stego-anal-simulator from scratch but to understand how techniques can be developed to detect steganography.

Steganography on Digital Images
Steganography is the art of hiding information in ways that prevent the detection of the hidden messages.Eventhough there are various ways by which this method can be implemented ,this project deals with the implementation of Steganography on Windows Bitmap images using LSB insertion Technique with a novel approach of "masking". Masking reduces the complexity of operation and is also more efficient than the other bit-reversal or bitshifting techniques.A version of the software STEGO-SOFT was developed for hiding text into 320*200 Windows Bitmaps. Later, the system was developed to hide any kind of data any file) into the image; with the provision for automatic zooming for large size of data to be encrypted. A much more advanced and highly efficient technique for implementing the same is under consideration using wavelet transforms.