tinyML-person_detection

Experimenting with tinyML - person detection

image_viewer.py

script to save the values retrieved from a camera module OV7675 as .jpg

camera_cont_display.py

script to continuously display an image from a camera module OV7675 It also displays label and prediction.

person_detection.py

creates a neural network using tensorflow to detect a person on the image. It expects the dataset in a folder dataset with subfolders:

0 a dataset 60 images of various background

1 a dataset of 60 images of human faces

It expects test images in a folder test (used for inference) with the same subfolder setup.

Images retrieved from OV7675 camera saved with image_viewer.py script

The script saves a model to a file.

use_person_detection_model.py

Checks how the model performs when loaded from a file and when converted to tflite. It expects test images in a folder test (used for inference) Saves a model as c array.