Get Started
Home
Topics
Search
Library
Research questionHow can point-supervised change detection produce reliable pixel-level changes from sparse annotations?Sparse point annotations cover only a small portion of each image, so the pixel-level labels needed for change detection are often incomplete and noisy. The resulting pseudo-label errors can obscure change boundaries and propagate through training.
AI
Computer Vision
Image & Video Processing
Machine Learning
Research Paper
Latest papersRecent research connected to this question, newest first.Progressive Pseudo-Label Optimization for Point-Supervised Change DetectionThe source addresses point-supervised change detection using bi-temporal images. Its proposed framework uses SAM2-generated candidate masks, bi-temporal mask selection, uncertainty-aware CNN refinement, and teacher-student self-training with periodically refreshed pseudo-labels. Evidence comes from experiments on WHU-CD, LEVIR-CD, and SYSU-CD, where it outperformed prior weakly supervised approaches on most benchmarks and remained competitive with some fully supervised methods.research paper · Sep 2, 2026
Related questions
How can unsupervised remote sensing detect unknown, complex changes without labeled examples?How can plant point-cloud segmentation adapt across species and sensing conditions with few labeled examples?How can weakly supervised video anomaly detection adapt to anomalies with varying durations and temporal dynamics?How can event-camera detectors encode temporal information without recurrence under rapid scene changes and difficult illumination?