Showing papers by "Malay K. Kundu published in 2007"

Robust SS watermarking with improved capacity

Abstract: Robustness is a key attribute of spread spectrum (SS) watermarking scheme. It is significantly deteriorated if one tries to achieve high embedding rate keeping other parameters unaltered. In literatures, typically various transformations like DFT, DCT, Fourier-Mellin and wavelet are used for SS multimedia watermarking but little studies have been attempted so far to see what are the possible factors which can improve robustness. The current paper has critically analyzed few such factors namely design of code pattern, proper signal decomposition suitable for data embedding, direction of decomposition, selection of regions for data embedding, signaling scheme, choice of modulation functions and embedding strength. Based on the observation, wavelet based SS watermarking scheme is proposed and improvement in robustness performance is verified through experimental results as well by mathematical analysis.

Image Segmentation Using Wavelet Packet Frames and Neuro-fuzzy Tools

Abstract: The present article describes a image segmenta- tion technique using M -band wavelet packet frames features. Those wavelet features are evaluated and selected using an efficient neuro-fuzzy feature evaluation techniqu. Both the feature extraction and neuro-fuzzy feature evaluation processes are of unsupervised type. They do not require the knowledge of number and distribution of classes corresponding to various landcovers in remotely sensed images. The effectiveness of the methodology is demonstrated on IRS-1A and SPOT images. Copyright c ∞ 2007 Yang's Scientific Research Institute, LLC. All rights reserved. Index Terms—M -band wavelet packet frames, adaptive basis selection, fuzzy feature evaluation index, neural networks, re- motely sensed image

A Co-processor for Computing the Euler Number of a Binary Image using Divide-and-Conquer Strategy

Abstract: Euler number is a fundamental topological feature of an image. The efficiency of computation of topological features of an image is critical for many digital imaging applications such as image matching, database retrieval, and computer vision that require real time response. In this paper, a novel algorithm for computing the Euler number of a binary image based on divide-andconquer paradigm, is proposed, which outperforms significantly the conventional techniques used in image processing tools. The algorithm can be easily parallelized for computing the Euler number of an N × N image in O(N) time, with O(N) processors. Using a simple architecture, the proposed method can be implemented as a special purpose VLSI chip to be used as a co-processor.

Image retrieval using fuzzy relevance feedback and validation with MPEG-7 content descriptors

Abstract: Content-Based Image retrieval has emerged as one of the most active research directions in the past few years. In CBIR, selection of desired images from a collection is made by measuring similarities between the extracted features. It is hard to determine the suitable weighting factors of various features for optimal retrieval when multiple features are used. In this paper, we propose a relevance feedback frame work, which evaluates the features, from fuzzy entropy based feature evaluation index (FEI) for optimal retrieval by considering both the relevant as well as irrelevant set of the retrieved images marked by the users. The results obtained using our algorithm have been compared with the agreed upon standards for visual content descriptors of MPEG-7 core experiments.

Image search using distance measurement method

Abstract: PROBLEM TO BE SOLVED: To provide an effective search method that uses image features. SOLUTION: An inquiry image 30q is identified from among a plurality of images in a database 40, using a Mahalanobis distance measurement section 60. This measurement section 60 is used for arraying the similarity of one image or more with respect to the inquiry image 30q. For all the images in the database 40, dispersion covariance matrices are calculated. Using this dispersion covariance matrices, the Mahalanobis distances between the inquiry image 30q and one or more images in the database 40 is calculated, and a range tree for identifying hopeful image candidates for measuring the Mahalanobis distance can be used. COPYRIGHT: (C)2007,JPO&INPIT