Attentioned Convolutional LSTM InpaintingNetwork for Anomaly Detection in Videos
This work addresses anomaly detection in videos, which is important for applications like surveillance, but it appears incremental as it builds on existing methods with specific modifications.
The authors tackled anomaly detection in videos by proposing a semi-supervised model that extends the Video Pixel Network with a convolutional attention mechanism and modifies the decoder for frame inpainting, using reconstruction error as an anomaly indicator; they tested it on a modified moving MNIST dataset and found it effective.
We propose a semi-supervised model for detecting anomalies in videos inspiredby the Video Pixel Network [van den Oord et al., 2016]. VPN is a probabilisticgenerative model based on a deep neural network that estimates the discrete jointdistribution of raw pixels in video frames. Our model extends the Convolutional-LSTM video encoder part of the VPN with a novel convolutional based attentionmechanism. We also modify the Pixel-CNN decoder part of the VPN to a frameinpainting task where a partially masked version of the frame to predict is given asinput. The frame reconstruction error is used as an anomaly indicator. We test ourmodel on a modified version of the moving mnist dataset [Srivastava et al., 2015]. Our model is shown to be effective in detecting anomalies in videos. This approachcould be a component in applications requiring visual common sense.