Video Segmentation Using Short Term Hierarchical Fast Watershed Algorithm
R. Durga, G. Yamuna, R. Barkavi · 2018
In computer vision, the most fundamental problem is the Segmentation, though more researches are carried out in this field, it is still very challenging. Segmentation is broadly classified into two broad categories, such as Image Segmentation and Video Segmentation. Comparatively video segmentation has a more general objective than image segmentation due to its large storage size and having temporal information. Hence, a Short Term Hierarchical Fast Watershed Algorithm (STHFWA) has been proposed to achieve an early segmentation by means of K-means clustering where the frames are segmented first and then the key frames into final segments by using Fast Watershed Algorithm (FWA). Finally, the Dual Tree Complex Wavelet Transform (DTCWT) is used. With these proposed algorithms, the objects can be successfully extracted.