Photo credit: MIT
Researchers from MIT, the MIT-IBM Watson AI Lab and other establishments have developed an AI model that dramatically speeds up high-resolution computer vision by reducing the computational complexity of the task. This model is capable of performing semantic segmentation – a deep learning algorithm that associates a label with every pixel in an image – accurately in real-time on a device with limited hardware resource, like computers found on autonomous vehicles.
The new model series for high-resolution computer vision performs up to nine times faster than its predecessors when deployed on a mobile device, while exhibiting the same or better accuracy than alternatives. Aside from heping autonomous vehicles make decisions in real-time, the model could also improve the efficiency of other high-resolution computer vision tasks including medical image segmentation.
- NO WIRES, MORE FUN — Break free from cords. Game, play, exercise and explore immersive worlds — untethered and without limits.
- 2X GRAPHICAL PROCESSING POWER — Enjoy lightning-fast load times and next-gen graphics for smooth gaming powered by the SnapdragonTM XR2 Gen 2...
- EXPERIENCE VIRTUAL REALITY — Take gaming to a new level and blend virtual objects with your physical space to experience two worlds at once.
While researchers have been using traditional vision transformers for quite a long time, and they give amazing results, we want people to also pay attention to the efficiency aspect of these models. Our work shows that it is possible to drastically reduce the computation so this real-time image segmentation can happen locally on a device,” said Song Han, an associate professor in the Department of Electrical Engineering and Computer Science (EECS).
