Abnormality Detection Using MODTA Scheme

Crowd analysis has become trending topic in Computer Vision field. There has been tremendous demand for smarter video surveillancefor private as well as public space. While considering the crowded scene it is required to have prior information and it is very difficult to extract from the complex video scene. The other major reason behind this is cost. Hence, in this paper we have presented MoDTA(multi-observation detection and tracking approach) based on the observational filter. Initially MoDTA acquires the location of people in given images, this helps in detecting the conviction value at pointed location. Later in tracking phase, MoDTA computes MO(Multiple Observed) weighted values along with the individual features. During motion, advection particle model has been used to facilitate dense scenario ranking. Correlation coefficient is used as a template detector and in order to find the estimation of theupcoming entity, template detector is used. At last in order to evaluate the system performance proposed model MoDTA is compared with respect to the Ground Truth.