Hierarchical Video Segmentation Using an Observation Scale

Hierarchical video segmentation provides region-oriented scale-space, i.e., a set of video segmentations at different detail levels in which the segmentations at finer levels are nested with respect to those at coarser levels. Hierarchical methods have the interesting property of preserving spatial and neighboring information among segmented regions. Here, we transform the hierarchical video segmentation into a graph partitioning problem in which each part will correspond to one region of the video. Thus, we propose a new methodology for hierarchical video segmentation which computes a hierarchy of partitions by a reweighting of original graph in which a segmentation can be easily infered. The temporal coherence is given, only, by color information instead of more complex features. We provide an extensive comparative analysis, considering both quantitative and qualitative assessments showing efficiency, ease of use, and temporal coherence of our methods. According to our experiments, the hierarchy infered by our two methods, p-HOScale and cp-HOScale, produces good quantitative and qualitative results when applied to video segmentation. Moreover, unlike other tested methods, our methods are not influenced by the number of supervoxels to be computed, as shown in the experimental analysis, and present a low space cost.