Search

Colon cancer is one of the leading cancers in both men and women across the globe. The incidence is especially high in developed countries and urban settings. Colonoscopy has been the established procedure for screening and early detection of colon cancer. This procedure involves insertion of a colonoscope with a video camera to examine the inner lining of the colon for suspicious polyps and growths which may develop into cancer. This requires two days of elaborate patient preparation to cleanse the colon of any residual matter which may obstruct the procedure. The colon is also air distended to expand it for clear access. The actual procedure itself takes a few hours and usually requires some kind of sedation to reduce the discomfort, especially to elderly patients. Large polyps and improper preparation also may prevent the procedure from completion, requiring repeat procedures. Growths on the outer surface of the colon also go completely undetected. Virtual colonoscopy is emerging as an alternative to conventional colonoscopy. This is simpler, less intrusive and faster. It is not affected by physical blockades and enables visualization of both inner and outer surfaces of the colon. An intrusive procedure is required only if suspicious polyps are located. This paper presents the methods and systems for developing an affordable virtual colonoscopy workstation which can be expanded to other endoscopic procedures also. It was evaluated on 21 independent colon data sets.

Compressed sensing is a revolutionary paradigm in data acquisition and processing, having impact in several fields. In the complex environment of ocean medium, where early detection of underwater targets is vital, compressed sensing assumes great significance as it brings down the computational resource requirements by reducing the data rates and hardware.In this paper DOA estimation is attempted with a linear array targeting underwater vessels. Array processing is investigated with compressed sensing based techniques and adaptive beam forming. Sparse reconstruction models are developed in this context for signal recovery. Finally the algorithms are validated on a real data set collected from an ocean experiment.

In this paper, the problem of direction of arrival estimation of unknown sources in under water scenario, withminimum data acquisition is investigated. The scheme is based onCompressive Sampling, an emerging technique in signal processing, which asserts that data acquisition and signal reconstructionis possible with much less number of measurements. The sparsenature of the angle spectrum is effectively utilized. The targetdata as received by a linear array of acoustic sensors is simulatedand compressed sensing is employed prior to processing. For reconstruction of compressively sampled signals, greedyalgorithms like orthogonal matching pursuit, matching pursuitand weak matching pursuit are employed. The performance ofthese algorithms are assessed and results are compared withconventional MUSIC algorithm.