BEM is a training-free inference module that suppresses false positives in fixed-background object detection by maintaining background embedding prototypes and applying an inverse-similarity rank-weighted penalty to detection logits.
Class imbalance in object detection: An experimental diagnosis and study of mitigation strategies
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
Knowledge distillation trains a 3.9x smaller YOLO student to retain 14.5% higher precision than direct training under INT8 quantization on BDD100K, exceeding the large teacher's FP32 precision while cutting false alarms.
citing papers explorer
-
BEM: Training-Free Background Embedding Memory for False-Positive Suppression in Real-Time Fixed-Background Camera
BEM is a training-free inference module that suppresses false positives in fixed-background object detection by maintaining background embedding prototypes and applying an inverse-similarity rank-weighted penalty to detection logits.
-
Edge AI for Automotive Vulnerable Road User Safety: Deployable Detection via Knowledge Distillation
Knowledge distillation trains a 3.9x smaller YOLO student to retain 14.5% higher precision than direct training under INT8 quantization on BDD100K, exceeding the large teacher's FP32 precision while cutting false alarms.