Today, I will write the tutorial to develop mask face detection using mmdetection. Firstly, we must collect the mask face data and lable it. Because the mmdetection use COCO format so after data is labeled, we must convert data to COCO format
Preapare dataset
- We can download the labeled data (https://github.com/AIZOOTech/FaceMaskDetection)or raw data (WIDER Face or MAFA), then annotate it.
- We use labelImg (https://github.com/tzutalin/labelImg) to annotate.
- Set
class nameismaskandno mask - We will get the xml file corresponding
- Set
Install mmdetection
- Firstly, we will install conda to handle environment for project (https://www.digitalocean.com/community/tutorials/how-to-install-anaconda-on-ubuntu-18-04-quickstart)
- We can follow this post to install mmdetection (https://github.com/open-mmlab/mmdetection/blob/master/docs/get_started.md)