Learning random-walk label propagation for weakly-supervised semantic segmentation

Abstract

Large-scale training for semantic segmentation is challenging due to the expense of obtaining training data for this task relative to other vision tasks. We propose a novel training approach to address this difficulty. Given cheaply-obtained sparse image labelings, we propagate the sparse labels to produce guessed dense labelings. A standard CNN-based segmentation network is trained to mimic these labelings. The label-propagation process is defined via random-walk hitting probabilities, which leads to a differentiable parameterization with uncertainty estimates that are incorporated into our loss. We show that by learning the label-propagator jointly with the segmentation predictor, we are able to effectively learn semantic edges given no direct edge supervision. Experiments also show that training a segmentation network in this way outperforms the naive approach.

arXiv commentThis is a revised version of a paper presented at CVPR 2017 that corrects some equations. See footnotes

Cited by in corpus (13)

• Affinity Attention Graph Neural Network for Weakly Supervised Semantic Segmentation
• CLIP in Medical Imaging: A Survey
• Saliency Guided Inter- and Intra-Class Relation Constraints for Weakly Supervised Semantic Segmentation
• Gaussian Dynamic Convolution for Efficient Single-Image Segmentation
• Learning from Pixel-Level Label Noise: A New Perspective for Semi-Supervised Semantic Segmentation
• Point Label Aware Superpixels for Multi-species Segmentation of Underwater Imagery
• Smooth Adversarial Examples
• Learning structure-aware semantic segmentation with image-level supervision
• Importance Sampling CAMs for Weakly-Supervised Segmentation
• Weakly Supervised Semantic Segmentation Using Constrained Dominant Sets
• DisFaceRep: Representation Disentanglement for Co-occurring Facial Components in Weakly Supervised Face Parsing
• Learning with Weak Annotations for Robust Maritime Obstacle Detection
• High-fidelity Pseudo-labels for Boosting Weakly-Supervised Segmentation