Segmentation in Video Image Sequences Using Seeded Region Growing

摘要

Traffic surveillance is the most important area of research which involves the safety issues in driving. Video Sequences are taken and it is converted as image sequences. By calculating the centroid of the image the image sequences are analyzed. The centroid value has been chosen as seed point in the image. From the seed point the region is expanded using the seeded region growing algorithm. Similarities of the pixels are considered within neighborhood pixels and based on the threshold value the segmentation have been done to identify the object. Using the seeded growing method the efficiency of the segmentation in image sequences is improved.