Detection of abnormal behaviour in video objects

A camera system for classifying a behavior of a detected object in a video frame, said camera system comprising:

• a lens system for focusing on a scene including the detected object;
• a camera module coupled to the lens system to store the video frame;
• an object detection module for detecting an object in the video frame, wherein the video frame comprises a plurality of blocks and the detected object is associated with a subset of the plurality of blocks;
• an object tracking module for associating parameters of a set of parameters with the detected object;
• and
• at least a processor and memory, cooperating to function as:
• an association unit configured to associate a behavior model with each of the plurality of blocks, wherein each behavior model includes a set of behavior modes having the set of parameters;
• a performance unit configured to, for each block in the video frame associated with the detected object:
• (i) determine a set of behavior statistics associated with the detected object, based on the detected object and the set of parameters;
• (ii) compare the determined set of behavior statistics with each behavior mode in the set of behavior modes associated with that block to determine an abnormality score;
• and
• (iii) select the lowest abnormality score from the determined abnormality scores;
• and
• a classifying unit configured to classify the behavior of the detected object based on the lowest abnormality scores associated with the subset of the plurality of blocks associated with the detected object, said classifying unit comprising:
• (i) a determining unit configured to determine an average abnormality score for all blocks in the subset of blocks associated with the detected object, based on the lowest abnormality scores;
• (ii) a comparing unit configured to compare the average abnormality score with an abnormality threshold;
• and
• (iii) a triggering unit configured to trigger an abnormality event if the average abnormality score exceeds the abnormality threshold.

A non-transitory computer-readable storage medium storing an executable computer program that when executed causes a computer to perform a method of classifying a behavior of a detected object in a video frame, the video frame comprising a plurality of blocks and the detected object being associated with a subset of the plurality of blocks, said method comprising steps of:

• (a) associating a behavior model with each of the plurality of blocks, wherein each behavior model includes a set of behavior modes having a set of parameters;
• (b) for each block in said video frame associated with the detected object:
• (i) determining a set of behavior statistics associated with the detected object in the video frame, based on the detected object and the set of parameters;
• (ii) comparing the determined set of behavior statistics with each behavior mode in the set of behavior modes associated with that block to determine an abnormality score;
• and
• (iii) selecting the lowest abnormality score from the determined abnormality scores;
• and
• (c) classifying the behavior of the detected object based on the lowest abnormality scores associated with the subset of the plurality of blocks associated with the detected object, said classifying step comprising:
• (i) determining an average abnormality score for all blocks in the subset of blocks associated with the detected object, based on the lowest abnormality scores:
• (ii) comparing the average abnormality score with an abnormality threshold;
• and
• (iii) triggering an abnormality event if the average abnormality score exceeds the abnormality threshold.