3 ms·
This board sets "the new standard for entry-level edge AI" at 30x the performance of the previous generation of "entry level". Now I can perform number plate re
by robga 4y ago
This board sets "the new standard for entry-level edge AI" at 30x the performance of the previous generation of "entry level". Now I can perform number plate recognition at 1000fps vs the previous 100fps. Is the previous board no longer OK for entry level tasks?
- DeWilde 4y agoEntry level tasks of yesteryear != entry level task of today.
- visarga 4y agoAt 100fps you can do superhuman agility robotics. This is a very old video, still impressive: https://youtu.be/-KxjVlaLBmk?t=155 https://youtu.be/-KxjVlaLBmk?t=155
- Roark66 4y agoGood luck getting low latency video into the board. The best previous jetson did I was told was 40ms. This is still 3~4 times better than any alternative, but nothing to write home about. Most delay happens in the isp path (image processing such as debayer, color correction, denoise, auto exposure). It seems many chips these days have the raw power to do those things quickly, but the the (chip)integrators don't really care. So then we get results like Rk3566 that can encode 10bit 720p in ~7ms, but the isp path adds 30~60ms of latency.