The uncorrelated frame-by-frame processing of videos not only affects the inference throughput adversely but also ignores the temporal information present in consecutive frames which drops the overall video segmentation accuracy.