Showing papers by "C. Krishna Mohan published in 2011"

De-duplication of photograph images using histogram refinement

Abstract: Content based image retrieval (CBIR), a technique which uses the content like color, texture and shape to search images from the large scale databases, is an active research area. In this paper, de-duplication process of photographs was implemented using CBIR. The CBIR technique uses color histogram refinement feature. The photograph data was divided into different clusters using k-means clustering algorithm. The clusters count depends on the numbers of photographs in each district of the state. The photo de-duplication exercise was carried out in a large photograph database which contains 22 million (approximately) photograph images. The experimental results shows that there were 0.35 million (approximately) duplicate photographs.

De-noising Slap Fingerprint Images for Accurate Slap Fingerprint Segmentation

Abstract: Fingerprints have unique properties like distinctiveness and persistence. Sometimes, fingerprint images can have some noisy data while capturing them using slap fingerprint scanners. This noise causes improper slap fingerprint segmentation due to which the performance of fingerprint matching decreases. The process of eliminating duplicates is called de-duplication which requires the plain quality fingerprints. While doing the segmentation of slap fingerprints, some of the fingerprint images are improperly segmented because of the noise present in the data. In this paper, an attempt is made to remove the noise present in the slap fingerprint data using binarization of slap fingerprint image, and region labeling of desired regions with 8-adjacency neighborhood for accurate slap fingerprint segmentation. Experimental results demonstrate that the fingerprint segmentation rate is improved from 78% to 99%.

ROI-based tissue type extraction and volume estimation in 3D brain anatomy

Abstract: In recent times, ROI-based extraction and volume estimation of various brain tissue types has gained immense attention from medical and computational research community. In diagnosing certain diseases, the volume of a specific brain region (e.g. Hippocampus) needs to be estimated to the best possible accuracy. Compared to the whole brain approaches, Region of Interest (ROI)-based approaches require an additional task of locating the ROI on subjects MRI. Given that no two brains are of the same size and shape, automatic extraction of ROI requires a certain degree of sophistication. The goal is to extract a particular region from brain imaging data and estimate the volumes of different tissue types (i.e., grey matter, white matter, and cerebro-spinal fluid) automatically. The experiments are conducted with ROI extraction using two softwares, namely, SPM and itk-SnAP. In the first one, whole brain segmentation is performed on normalized images and then a predefined mask is used to extract region of interest. itk-SnAP uses a region growing approach for segmentation. Results of these methods are compared and the analysis shows that both methods give comparable results. However, it is observed that itk-SnAP is more robust for calculating volumes of small regions. In this paper we proposed a pipelining process which contains different stages like normalization, ROI masks generation, segmentation and volume estimation by using VBM5 and itk-SnAP tools.