Hi, I am working on an image classification problem on Portenta. I got some results here but am also wondering how better an object detection application can do. Came across FOMO and am wondering:
- Is it open source? Is there a publication on its architecture?
- Is there a way to implement FOMO without Edge Impulse?
- What are the guidelines on the number of images (and annotations) per object required to build a good model?
PC