Decision-level fusion with WBF outperforms feature-level fusion for occlusion-robust detection on ultra-low-end hardware, with gains up to +0.3827 mAP across three views and on-device execution on Coral boards.
Image and Vision Computing107, 104117 (2021)
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
years
2026 3representative citing papers
ED-CCF projects detections into a quad-state error taxonomy and applies class-conditional calibration only when empirically justified, raising mAP50 for a hard class by 22.4% on a 600-image benchmark while keeping global mAP50 stable.