YoloV2 on multiple NCS

[idata]

A quick shout out to PINTO, who built this cool "multi-motion detection" system using TinyYoloV2, Raspberry PI and Intel Movidius Neural Compute Stick.

https://qiita.com/PINTO/items/7f13fcb7c894c27691b2

[idata]

@AshwinVijayakumar

Thank you for introducing my blog.

The next post will show twice or more performance despite being a full screen display with one stick.I tried combination with OpenGL.We hope for your reference.

https://qiita.com/PINTO/items/db3ab44a3e2bcd87f2d8

Thank you.

[idata]

@PINTO, happy to highlight a fellow developer's work! It might be worth checking out MobileNet SSD as well - https://github.com/movidius/ncappzoo/tree/master/caffe/SSD_MobileNet

[idata]

@AshwinVijayakumar

Thank you for giving me advice.I feel like I heard that SSD is later than YoloV2, but I am interested and I will implement it between houseworks.

[idata]

@AshwinVijayakumar

MobileNet SSD … I was surprised so greatly!!

It is twice the performance of TinyYolo.

Performance is too high, I can no longer understand the significance of multi stick.

Look at the following. [MobileNet SSD + Neural Compute Stick + Raspberry Pi with Multi Stick / OpenGL]

https://qiita.com/PINTO/items/b97b3334ed452cb555e2

[idata]

Withdraw the previous word.

Three sticks made it even faster.https://youtu.be/sQnFbRSqIA8

[idata]

@PINTO Nice! Thanks for sharing.