Head Detection Using YOLO Algorithm

The objective is to train a YOLO algorithm to detect multiple heads from a frame.

Getting Started

Prerequisites

1. TensorFlow
2. Keras

Download the pre-trained weights for the backend

Download full_yolo_backend.h5 from https://drive.google.com/file/d/1Q9WhhRlqQbA4jgBkCDrynvgquRXZA_f8/view?usp=sharing Put it in the root directory.

Dataset

Download the dataset and put it in the root directory.

Images - https://drive.google.com/open?id=1zn-AGmsBqVheFPnDTXWBpeo3XRH1Ho15

Annotations - https://drive.google.com/open?id=1LiTDMWk0KglGueJCaxgneEA_ltvEbUDV

Training

Run train.py

The weights for the front-end will be saved in the file name "model.h5".

Inference

I have uploaded the front-end weights. https://drive.google.com/file/d/1wg4q9cc6q04oRr_Xaf9GVhhbTSH4-ena/view?usp=sharing

Give the path to your image in predict.py

Run predict.py

License

This project is licensed under the MIT License