I have an OpenMV M7 camera and have folowed the “Adding sight to your sensors” tutorial
I have a trained model with lots of images who performs great…
The image size is set to 48x48 / grayscale and I use MobileNetV2 0.35 (all others perform very bad)
On the OpenMV I get the
MemoryError: Out of fast Frame Buffer Stack Memory! Please reduce the resolution of the image you are running this algorithm on to bypass this issue!
Kwabena at OpenMV says it used to work, if I set the image to 48x48 in EI🤔
You probably need to use a smaller MobileNet as you’re limited in in RAM on the M7.
Here’s a tutorial with the H7 here (with 1MB RAM): https://www.hackster.io/ishotjr/mega-or-mini-image-classification-on-the-1mb-openmv-cam-h7-be57ac
Could you try selecting the MobileNet v2 0.1, or even 0.05 if needed?
Or one of the new MobileNet v1 models, which are smaller too and will work with larger images (so you don’t lose too much detail).
I finally managed to make it work Thank you for your help…