Coarse-to-Fine Feature Mining for Video Semantic Segmentation.
Guolei SunYun LiuHenghui DingThomas ProbstLuc Van GoolPublished in: CoRR (2022)
Keyphrases
- coarse to fine
- semantic segmentation
- multiscale
- superpixels
- object detection
- multiresolution
- object class
- conditional random fields
- image registration
- object categories
- video sequences
- weakly supervised
- scene classification
- image features
- feature vectors
- dynamic programming
- video frames
- pixel level
- markov random field
- generative model
- machine learning
- denoising
- object recognition
- image processing