Tensorflow & Opencv

[salarizadi]

Hello to all.

It would be great if we could add opencv and tensorflow to moddable. I am researching this, trying to figure out how to do it, but with your help we will add more opencv, tensorflow and AI to moddable. Let me know if you have any ideas.

[phoddie]

What are you thinking to use as a starting point?

[phoddie]

We have done some work on ESP32 camera a while back. It is unpublished because it felt very unstable. We could make it work with very specific configurations but nothing generally useful -- everything requires lots of tweaking. Maybe it has gotten better since then. Do you have any recent experience with ESP32 camera?

[salarizadi]

I will check this too, the esp32 camera is very important, we are going to make object detection bots from image and AI, but before that, if xs simulator can do the same for simulating the display for the camera , it will be great. Be ready to make the xs simulator camera so that you can make the necessary arrangements. Please let me know when you want to build the camera simulator

So that I can start the initial tests of object recognition from the image

[phoddie]

Thank you. I agree that simulator support will be important. We've done that kind of thing before. But.... the design of the Camera API needs to be driven by the embedded device (ESP32). So... we should understand how this will work on the embedded device first. That will then define how to integrate camera support into the simulator.

[salarizadi]

Please continue the discussion on Telegram, it is easier and faster. Telegram ID: izsalar

[phoddie]

I'm not a Telegram user. I could learn, but I very much prefer to keep technical conversations like this available to as many people as possible, so that they can learn and contribute too. Maybe our Embedded JavaScript Gitter would be a good to try for conversations?