hkchengrex/STCN
PyTorch implementation of Space-Time Correspondence Networks (STCN) for efficient video object segmentation achieving state-of-the-art results on DAVIS and YouTubeVOS benchmarks.
Not currently ranked — collecting fresh signals.
star history
STCN presents a framework for modeling space-time correspondences in video object segmentation. The approach uses neural networks to track and segment objects across video frames by learning correspondence relationships in space and time. Built with PyTorch, the model achieves competitive accuracy on benchmarks like DAVIS 2017 and YouTubeVOS while running at real-time speeds of 20+ FPS. This is a research implementation with code for training and evaluation.
Frequently asked
- What is hkchengrex/STCN?
- PyTorch implementation of Space-Time Correspondence Networks (STCN) for efficient video object segmentation achieving state-of-the-art results on DAVIS and YouTubeVOS benchmarks.
- Is STCN open source?
- Yes — hkchengrex/STCN is open source, released under the MIT license.
- What language is STCN written in?
- hkchengrex/STCN is primarily written in Python.
- How popular is STCN?
- hkchengrex/STCN has 568 stars on GitHub.
- Where can I find STCN?
- hkchengrex/STCN is on GitHub at https://github.com/hkchengrex/STCN.