Motion-Attentive Network for Detecting Abnormal Situationsin Surveillance Video
U-Ju Gim, Jeong-Hun Kim, Kwan‐Hee Yoo, Aziz Nasridinov · 2020
Recently, numerous studies have utilized deep-learning-based approaches to detect anomalies in surveillance cameras. However, while several of these studies used motion features to detect abnormal situations, detection problems can arise due to the sparse information and irregular patterns in certain abnormal situations. We propose a means of preserving motion patterns in abnormal situations through a network called MA-Net, which solves representation problems caused by a loss of sparse information and irregular patterns. We show through experiments that the proposed method is superior to state-of-the-art methods.