Anomaly Detection from Crowded Video by Convolutional Neural Network and Descriptors Algorithm: Survey

Depending on the context of interest, an anomaly is defined differently. In case when a video event isn't expected to take place in video, it seen as anomaly. It can be difficult describe uncommon events complicated scenes, but this problem frequently resolved by using high-dimensional features well descriptors. There difficulty creating reliable model trained with these descriptors because needs huge number training samples and computationally complex. Spatiotemporal changes or trajectories are typically represented that extracted. The presented work presents numerous investigations address issue abnormal detection from crowded its methodology. Through use low-level features, like global local feature features. For most accurate identification anomalous behavior videos, attempting compare various techniques, uses more dataset require light weight for diagnosing anomalies objects through recording tracking movements extracting features; thus, should strong differentiate objects. After reviewing previous works, noticed there need accuracy modeling decreased time, since attempted real-time outdoor scenes.