hustvl/SparseInst
A real-time instance segmentation framework based on sparse instance activation maps, achieving 37.9 AP at 40 FPS.
Not currently ranked — collecting fresh signals.
star history
SparseInst introduces Instance Activation Maps (IAM) to adaptively highlight informative object regions for recognition. It is a fully convolutional instance segmentation framework that eliminates the need for non-maximum suppression (NMS) or sorting, making it efficient and easy to deploy. The method achieves competitive accuracy-speed trade-offs on standard benchmarks like MS COCO.
Frequently asked
- What is hustvl/SparseInst?
- A real-time instance segmentation framework based on sparse instance activation maps, achieving 37.9 AP at 40 FPS.
- Is SparseInst open source?
- Yes — hustvl/SparseInst is open source, released under the MIT license.
- What language is SparseInst written in?
- hustvl/SparseInst is primarily written in Python.
- How popular is SparseInst?
- hustvl/SparseInst has 621 stars on GitHub.
- Where can I find SparseInst?
- hustvl/SparseInst is on GitHub at https://github.com/hustvl/SparseInst.