KIRAN KUMAR RAJENDRA KUMAR

This work implements Adaptive Noise Cancellation in Frequency domain, where the channel is estimated using adaptive filter and noise from the channel is cancelled to obtain a clean speech. This filter implements Block type LMS filter for adaptive algorithm and error convergence curves are plotted.

This project uses the Generalized Likelihood Ratio Test using MATLAB to estimate the unknown signal parameters of a given signal, here an HFM signal from Underwater Radar is used. There are two datasets given for reference out of which one contains the actual data. The mesh plot is used as the decision statistic to check whether the given signal satisfies the condition

This project is used for the estimation of channel present in OFDM through using Pilot Sequences (Training Sequences) and Channel taps. The performance of the system is compared using BER curves that are obtained for evenly spaced Pilots and closely spaced Pilots. Also, the number of channel taps are varied to estimate the best BER curve

This project involves studying various diversity combining techniques like Maximal Ratio Combining, Selection Combining, Equal Gain Combining, Post Detection Combining that are used in modulation schemes BPSK using MATLAB. By performing Monte-Carlo Simulations on MATLAB, the Average Bit Error Rate(BER) for each of the techniques are found and they are plotted in a logarithmic scale against Signal to Noise Ratio(SNR).

This work involves using feature detectors FAST and AGAST to detect the key points in two images obtained from two cameras which indicate left view and right view. FREAK and LUCID are used to calculate the descriptors for the key points that are found. By using thresholding and matching algorithms the key points in two images are matched and a disparity value is calculated for those key points. By using this disparity matrix, the depth map for the particular frame can be calculated. The dataset used here was 2006 dataset from Middlebury Stereo Evaluation.

App is useful for color blind people as they are able to point a object using their Android Smartphone camera and are able to know the true color of the object by touching the particular object in the frame. The camera preview is displayed in the app and when there is a touch registered using OnTouchListener, a green square appears along the touch area and which contains the true color as the title. Also using TTS the color is read out to the user. Open CV for Android is used for this project and using the HSV cone, the color is classified