Enhanced Object Detection and Tracking Algorithm in Complex Videos

An enhanced algorithm to track objects in complex video scenes is developed and presented in this paper. A single Gaussian method is proposed to detect the object and also a mean-shift color based algorithm is applied to achieve the appropriate tracking of the detected object. The centroid detection via k-mean clustering is implemented to reduce the number of bins, which is important to extract the N centroid for the clustering process. After producing the algorithm of the proposed method, system implementation has been done through the MATLAB and developed coding was practiced at different videos and results achieved as it is provided in testing part.