DICTIONARY DESIGN FOR COMPUTATIONALLY EFFICIENT VIDEO ANOMALY DETECTION VIA SPARSE RECONSTRUCTION TECHNIQUES

摘要

Methods, systems, and processor-readable media for pruning a training dictionary for use in detecting anomalous events from surveillance video. Training samples can be received, which correspond to normal events. A dictionary can then be constructed, which includes two or more classes of normal events from the training samples. Sparse codes are then generated for selected training samples with respect to the dictionary derived from the two or more classes of normal events. The size of the dictionary can then be reduced by removing redundant dictionary columns from the dictionary via analysis of the sparse codes. The dictionary is then optimized to yield a low reconstruction error and a high-interclass discriminability.

主权项

1. A method for pruning a training dictionary for use in detecting anomalous events from surveillance video, said method comprising:
receiving a plurality of training samples corresponding to normal events; constructing a dictionary comprising at least two classes of normal events from said plurality of training samples; generating a plurality of sparse codes for selected training samples among said plurality of training samples with respect to said dictionary derived from said at least two classes of normal events; reducing a size of said dictionary by removing redundant dictionary columns from said dictionary via analysis of said plurality of sparse codes; and optimizing said dictionary to yield a low reconstruction error and a high-interclass discriminability.