pjarbas / headphones-detection-yolo

Headphones object detection using YoloV4

Headphones object detection using YoloV4

Headphones Detection 1	Headphones Detection 2	Headphones Detection 3

This project aims to train a object detecion deep learning model, using a custom number of classes (Headphones)

Get started

Ensure that your python version is >= 3.9 and you have installed the opencv with CUDA

Download the data running the following command in the src directory:

   $ python data_downloader.py 

Create the train.txt and test.txt files with random split, these files must be in the same directory that the darknet.data file

   $ cd src
   $ python split_train_test.py ../data/images/ 

Install YoloV4 Darknet with CUDA:

• https://github.com/AlexeyAB/darknet#how-to-compile-on-windows-using-vcpkg

Download the yolov4-tiny-custom

Get the yolov4-tiny-custom.cfg and change the number of classes and the filters for the [conv] layers before the [yolo] layers

• check the documentation to see all the parameters to be changed: https://github.com/AlexeyAB/darknet#how-to-train-to-detect-your-custom-objects

   $ wget https://github.com/AlexeyAB/darknet/blob/master/cfg/yolov4-tiny-custom.cfg

Train using pre-trained weights

Download file with the first 29-convolutional layers of yolov4-tiny: https://github.com/AlexeyAB/darknet/releases/download/darknet_yolo_v4_pre/yolov4-tiny.conv.29

   $ darknet.exe detector train darknet.data yolo-cfg/yolov4-tiny-custom.cfg yolov4-tiny.conv.29 -map

Test the results

   $ darknet.exe detector test darknet.data yolo-cfg/yolov4-tiny-custom.cfg backup/yolov4-tiny-custom_final.weights data/your-image-path.png

Run Inference in python

   $ cd src/
   $ python main.py --image your_image.png