Gaussian mixture model and spatial-temporal evaluation for object detection and tracking in video surveillance system

Scene analysis is very important in a video surveillance system, with purpose to gain information and knowledge from the surrounding. There are many researches covering problems in object detection and tracking, but solve it only partially. This paper will cover an integral technique to do object detection and tracking for video surveillance. First, pixels in the images will be modelled with gaussian mixture model with K-Means algorithm to separate foreground from background image. Then, morphological cleaning is applied to remove noise pixels. Objects will be formed with spatial evaluation, with color mean and contour chain code as its feature. Tracking will be performed with temporal evaluation, i.e. inter-frame object features and distance comparison. This technique is doing well in object detection and tracking, with high true positive and low false negative, but still suffering from false positive in dynamic background scene. The implementation is not perfect, either, with only 30%-50% video speed from the original.