Fig. 6 | Scientific Reports

Fig. 6

From: Attention-guided saptio-temporal feature fusion for robus video surveillance anomaly detection

Fig. 6The alternative text for this image may have been generated using AI.

presents the proposed temporal confidence reweighting loss (TCR-Loss), which introduces temporal consistency into the detection confidence learning process. Instead of treating each frame independently, the model analyzes consecutive frames \(\:\left(\text{t}-1,\text{t},\text{t}+1\right)\) and penalizes abrupt variations in predicted confidence scores across time.

Back to article page