As we’ve been saying for a while, deep learning and artificial intelligence (AI) are moving into the mainstream in a big way. Highly efficient vision processors bring this intelligence into small, low-power devices.
In the embedded world, as you can see from our report from Embedded Vision Summit (EVS) 2017, these powerful technologies are taking center stage. We heard great reports from the Embedded Systems Conference (ESC) in Boston, as well. Unfortunately, we couldn’t make it this year, but we were lucky to still have a demo running there, thanks to Embedded.com Editor in Chief Clive Maxfield, aka “Max the Magnificent”. We lent him a development board with our CEVA-XM4 IP and a trained neural network for object recognition. This was made possible by the CDNN2 (CEVA Deep Neural Network) software framework, which was recently the recipient of the prestigious CEM Editor’s Choice Awards. Max set up a camera connected to the board, and conducted some tests with various images and objects. He said that the guys in his office were really impressed with the object recognition capabilities of the embedded system. So much so that he decided to use it as part of his discussion on artificial intelligence at the conference. It’s definitely better to read about it in his signature colorful tone, so here’s the link to Max’s post on Embedded.com. Also, we are lucky enough to have a video of the demonstration he gave, and we are thrilled to share it with you. Here’s the link: ESC Boston 2017 – Clive ‘Max’ Maxfield talks about new technologies.
Of course, seeing the video is not the same as seeing the demo in action. Max tells us he will be taking it with him to a few more conferences this year. He’ll be giving the Keynote at the “Electronics of Tomorrow” conference in Denmark at the beginning of November and his “Advanced Technologies” talk at ESC Minneapolis in November and ESC Silicon Valley in December. If you’re attending any of these, you should definitely stop by.
To find out more about harnessing the power of the AI and deep learning in embedded systems, click here.
To see more CEVA-powered AI, check out this Real-Time Object Recognition demo. For more, visit our YouTube channel, here.