2nd Half of Thesis project, implementation of MTCNN onto an ESP32 The board is an ESP-Cam from AI-Thinker with 4mb of PSRAM.
This code implements P-Net and R-Net on the ESP32 using weights from the MTCNN model trained in MTCNN-Leaf
The weights are translated by translateWeights2ESP_lib.py. This goes from Caffe's convention to the ESP-FACE library.
--> ESP-Camera
--> ESP-FACE
ESP32-Cam by AI-Thinker
FTDI Programming:
Current Output:
The current state of the model works and provides output, but cannot save the jpeg image to an SD Card. More details on this issue can be found here: Issue 4495
O-Net will be implemented in the near future.
A full explanation can be read in my thesis, please email me if you are interested in seeing it. As there were many steps involved I would be happy to help on any developments in the future, Please file an issue or send me an email.