Pixel-Level Contrastive Learning
A representation learning method that extends the contrastive learning paradigm to the pixel level, treating pixels belonging to the same keypoint as positive samples and pixels of different keypoints as negative samples, and introducing learnable keypoint prototypes as anchors so that the network distinguishes different keypoints in embedding space.
Timeline (last 90 days)
The study uses a general U-Net backbone and, through pixel-level contrastive learning, produces heatmaps with greater discriminative power than traditional heatmap regression.
Pixel-level contrastive learning treats pixels belonging to the same keypoint as positive samples and pixels of different keypoints as negative samples, and introduces learnable keypoint prototypes as anchors.
All Facts (2)
The study uses a general U-Net backbone and, through pixel-level contrastive learning, produces heatmaps with greater discriminative power than traditional heatmap regression.
50%UnverifiedPixel-level contrastive learning treats pixels belonging to the same keypoint as positive samples and pixels of different keypoints as negative samples, and introduces learnable keypoint prototypes as anchors.
50%