Symmetric Mask Wavelet Based Detection and Tracking of Moving Objects Using Variance Method

Fast and accurate object tracking is very important for real time applications like video surveillance, traffic monitoring etc. In most of the conventional object tracking methods environmental changes, memory requirement and computation speed are the major constraints. This paper proposes an efficient object tracking method to compensate for all these challenges. Here a Symmetric Mask based on Discrete Wavelet Transform (SMDWT) has been used in order to compensate for illumination changes and low memory requirement. Symmetric Mask Discrete Wavelet Transform is a low complexity DWT where each filter is derived from the 2- D DWT of 5/3 integer lifting-based coefficients. For fast object tracking variance method is adopted where maximum nonzero pixel value is considered. The experimental results show that the proposed method yields better result on the basis of computational time, memory requirement, speed of operation and environmental changes thanthe conventional DWT based approach.