cv-mask-detection

Training custom object detector using yolo v3 and v4, along with inference and deployment.

This project will using Kaggle image set: Face Mask Dataset (YOLO format), for training object detector. https://www.kaggle.com/aditya276/face-mask-dataset-yolo-format

1. Change Runtime Type

Runtime -> Change runtime type -> Select hardware Accelerator as GPU.

2. Link Google Drive

from google.colab import drive
drive.mount('/content/drive')

Enter into the folder:
%cd /content/drive/My\ Drive/Mask\ Detector

3. Clone DarkNet Repository

!git clone https://github.com/AlexeyAB/darknet.git

4. Compile DarkNet

1. Build darknet with OpenCV
2. Build with CUDA enabled
3. Build with cuDNN enabled.

Enter into darknet folder:
%cd darknet

!sed -i 's/OPENCV=0/OPENCV=1/' Makefile

!sed -i 's/GPU=0/GPU=1/' Makefile

!sed -i 's/CUDNN=0/CUDNN=1/' Makefile

print("Building. . . It might take 2-3 minutes")

!make &> build_log.txt

print("Build Done!")

5. Prepare Dataset Files

see data_prepare.py. (This file modifies the .txt file directing the images and labels).

6. Prepare Configuration File for YOLO Training

1. train.cfg
2. test.cfg
3. setup.data
4. class.names

Should modify each file accordingly. See more details in appendix.

7. Prepare weights for Convolutional backbone

Yolov3 pretrained weights:
https://pjreddie.com/media/files/darknet53.conv.74

Yolov4 pretrained weights:
https://github.com/AlexeyAB/darknet/releases/download/darknet_yolo_v3_optimal/yolov4.weights

8. Start Training

For starting training using darknet, we need to execute the following command. Here we are specifying the:

1. path to the setup file,
2. path to config file,
3. path to convolutional weights file.
   !./darknet detector train ../configuration_v3/setup.data ../configuration_v3/train.cfg ./backup/darknet53.conv.74 -dont_show -map 2> train_log.txt

Experiments & Results