A new image-processing technology has been developed by Fujitsu, who claims that it now can be used to track people even in heavily blurred footage from security cameras. The company states that this tech is the first of its kind that can detect people in low-resolution imagery where faces are indistinguishable.
The technology is said to use computer-vision algorithms in order to analyse the footage and recognize shapes such as head and torso, which remain even if the image is distorted or multiple people in a frame overlap. The algorithm is then compared with footage from different cameras and determines if an individual is the same person by focusing on distinctive colours of a person’s clothing.
Fujitsu said that an indoor test of the system has been performed, having been able to track the paths of 80% of the test subjects. The company also claimed that detecting people’s movement could be useful for retail design, reducing pedestrian congestion in crowded urban areas or improving evacuation routes for emergencies.
Thank you PC World for providing us with this information