Surveillance cameras have become widespread. Many videos are recorded for watch-over and monitoring purposes, increasing security.
The increase in the number of videos and their lengths has a troublesome aspect. The amount of information contained in videos has rapidly grown, making it increasingly difficult for people to find what they should pay attention to.
Ricoh has developed a technology to extract unusual things and behaviors from videos.
Fan movement is not detected as an anomaly. A person entering the room and a PC left behind are detected as anomalies.
The usual lying, waking, and turning over are not detected as anomalies. Behaviors that differ from normal, such as unusual lying postures and a fall from the bed, are detected as anomalies.
Recently, much effort has been put into research and development of artificial intelligence (AI), enabling machines to have a learning function.
Generally speaking, to allow AI to learn something, you need to prepare everything beforehand. With anomaly detection, you need to identify and provide all abnormal states that should be detected.
Yet most of the videos recorded on surveillance cameras are of normalcy. Beyond that, it is unrealistic to predict every single anomaly in the first place.
Thus, Ricoh has come up with the idea of using semi-supervised anomaly detection, a technology proven in appearance inspection and frozen road detection. Anomalies can now be detected even when very little anomaly data is available.
The technique uses only the videos of normalcy for learning.
All scenes that depart from normalcy (for instance, an increase or decrease in things, or unusual behaviors) are determined as anomalies. This scheme allows the detection of unknown, different anomalies.
With this technique, all normalcy videos become sample videos of no anomalies. This allows many normalcy videos to be obtained easily.
Ricoh will continue to exploit its technological resources, researching systems that precisely capture the surrounding conditions and generate feedback from them.