FACE MASK DETECT AND DISTANCE AND CROWD DENSITY ESTIMATES

Introduction

This is major project of computer engineering about face mask detect and distance, crowd density estimates

Dev

Member:
 - DAO DUY NGU
 - LE VAN THIEN
Instructor: TRAN THI MINH HANH

Usage

Install package

git clone https://github.com/DuyNguDao/Project-Covid19.git
cd Project-Covid19
conda create --name covid19 python=3.8
pip install -r requirements.txt

Quick start

python distance_bird_eyes_video.py

Guides setup distance and Crowd density estimates

- Config for camera at bird_eyes_view.yalm
- At the first frame:
    + Click seven point:
      * click first four point: left, top right, bottom right, bottom left
      * click again three point then compute pixel width, height: top left, top right, bottom left

Datasets and result model training

dataset and result

Result confusion matrix on set of YOLOv5

Plot train

Train

Test

Compare between model YOLOv5 and RetinaFace + MobileNetV2

Confusion matrix of YOLOv5

Confusion matrix of RetinaFace + MobileNetV2

Result realtime

• Frame 1
• Frame 2
• Frame 3

Contact

email: ddngu0110@gmail.com