The human behaviour detection is performed with image pattern recognition system and method, identification of persons in the image, spatio-temporal saliency of motion.
Video access system and method based on “action type” detection. The system is capable of recognizing human actions when they occur in a video stream.
Regarding Human Behaviour, for specific SW libraries, Nurjana Technologies works with TNO (Organisation for Applied Scientific Research). Below a list of patents and a track record,
EP2410467, EP20100170682, WO2012011817: Identification of persons in the image. Topic: characterising persons based on appearance and matching between their occurrences in different images.
P99440PC00 Detection of human actions from video data. Topic: spatio-temporal saliency of motion. Some motion patterns in the video are more characteristic of particular human actions than other patterns.
P100091PC00 Image pattern recognition system and method. Topic: soft-assignment random forest. It classifies samples into target vs. non-target class. In our case, we use it for recognition of human actions in video streams..
P99622EP00 Video access system and method based on action type detection. Topic: specification of behaviour by rules. Some human actions can be described by humans in terms of rules. TNO has designed such a rule-based system, which is capable of recognizing human actions when they occur in a video stream.