Today, modern systems perceive and process the world around themThey're not good enough.
Two studies, from Bristol and Manchester, have shown how perception and learning can be combined to create new camerasThey presented a convolutional neural network (CNN) in the SCAMP-5D vision system.It is able to classify what is happening in front of it at a speed of 8,200 frames per second.
A convolutional neural network (CNN) in the SCAMP-5D computer vision system that classifies hand gestures at 8,200 frames per second. Credit: University of Bristol
Recognize images without the need to record them or send them further to processing systems.
SCAMP-5d technical vision system. Credit: University of Manchester
This approach will make the systems much more efficient and safer as there is no need to record images to analyze them.
Work made possible by architectureSCAMP, developed by Peter Dudek, professor at the University of Manchester, and his team. SCAMP is a camera processor chip, which scientists describe as a Pixel Processor Array (PPA). The PPA has a processor built into each pixel that can communicate with each other to process data in parallel. This is perfect for CNN and vision algorithms.
The development of scientists is presented at the European Conference on Computer Vision (ECCV).
Read more
See how the moon appeared. Ancient planet crashed into Earth
Archaeologists have found an ancient burial in the Crimea. There was a "ticket" to the afterlife
Abortion and science: what will happen to the children who will give birth