Entertainment industry is roofed with audio, video, graphics and multimedia. It deals with large variety of data. As the data increases, it becomes a troublesome problem especially in large database or realtime and telepresence applications, where the memory, bandwidth and storages are limited. So, in this study we propose, a novel compression method, which is achieved by video content analysis using Repeated Object Clips [ROC]. To achieve ROC, the key frame selection algorithm is used. In key frame selection process, we extract key frame from total number of frames. For selection process, initially frame separation process is carried out, then object detection and object segmentation is performed by level set method. By this novel technique moving object was tracked effectively and compression ratio increased significantly.