Fig. 6From: Deep anomaly detection through visual attention in surveillance videosPre-trained C3D feature extraction of obtained attention regionsBack to article page