Multiple object tracking by employing shaped-based features and Kalman filter
Felix Mattathil Philip, Rajeswari Mukesh · International Journal of Business Intelligence and Data Mining · 2017
There has been fast development happening in the multimedia and the related technologies, particularly associated with visual tracking and search operations. Moving target detection has been comprehensively engaged in various arenas but has the disadvantage that the scheme is frequently complex and also that tracking is affected numerous external factors. In this article, multiple objects recognition and tracking is projected so as to progress the method and make it more robust and general with assistance of shape-based features and Kalman filter. Primarily, the input video is rehabilitated to frames and then manually segmented for object segmentation. Consequently, the objects are tracked with the help of Kalman filtering. The method is assessed under standard evaluation metrics of error value and the score value. The technique achieved maximum score values of 95% and minimum error value of 25%. The results validate the effectiveness of the technique.