I have an OpenMV Cam H7 with which I want to run a tflite model.
My model is trained to work on 28x28x3 BGR images which are normalized to 0-1 by dividing each pixel by 255 at the time of training.
If I run this model on the board by setting the sensor.set_pixformat(sensor.RGB565) and give this image to tf.classify will ti work?
- Will it get resized to 28x28.
- Will the pixel values be normalized to 0-1 by OpenMV.
- Will its color space be converted to BGR
- As RGB565 has two channels and BGR has 3 channels.
Looking forward to hearing from you.
Thanks and Regards