Faster segmentation of objects in video with new affinity formula
--
Faster segmentation of objects in video with new affinity formula
Faster segmentation of objects in video with new affinity formula
arXiv paper abstract https://arxiv.org/abs/2106.05210
arXiv PDF paper https://arxiv.org/pdf/2106.05210.pdf
Project Web page https://hkchengrex.github.io/STCN
GitHub https://github.com/hkchengrex/STCN
… simple … approach to modeling space-time correspondences in … video object segmentation.
… establish correspondences directly between frames without re-encoding the mask features for every object
… propose using the negative squared Euclidean distance instead to compute the affinities.
… beneficial to both memory efficiency and inference accuracy.
… new state-of-the-art results on both DAVIS and YouTubeVOS datasets while running significantly faster at 20+ FPS for multiple objects without bells and whistles.
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
LinkedIn https://www.linkedin.com/in/morris-lee-47877b7b
