arghadeep25 / face-detection

Face Detection using CNN, HoG and SSD

Face Detection

Face detection is one the hot topics in the field of image processing. The objective is to find whether there is a face in an image or not. Various kind of approaches are taken so far to solve this problem. Some of the famous detection algorithms were discussed here.

Models

Haar Cascade

Source: https://www.cs.cmu.edu/~efros/courses/LBMV07/Papers/viola-cvpr-01.pdf

CNN (Convolutional Neura Network)

Source: https://arxiv.org/pdf/1502.00046.pdf

HoG (Histogram of Oriented Gradients)

Source: https://lear.inrialpes.fr/people/triggs/pubs/Dalal-cvpr05.pdf

SSD (Single Shot Detector)

Source: https://www.cs.unc.edu/~wliu/papers/ssd.pdf

Dependency

• Python 3.5
• Dlib (http://dlib.net/)
• Numpy (https://www.numpy.org/)
• OpenCV (https://opencv.org/)

Installation

pip3 install -r requirements.txt

Usage

User Guide

For help, you can use

python3 main.py -h   

For applying different models on images, you can use the following methods:

python3 main.py -m ssd -t image -i dataset/images/[IMAGE_NAME]
python3 main.py -m hog -t image -i dataset/images/[IMAGE_NAME]
python3 main.py -m haar -t image -i dataset/images/[IMAGE_NAME]
python3 main.py -m cnn -t image -i dataset/images/[IMAGE_NAME]

For applying different models on videos, you can use the following methods:

python3 main.py -m ssd -t video -i dataset/video/[video_NAME]
python3 main.py -m hog -t video -i dataset/video/[video_NAME]
python3 main.py -m haar -t video -i dataset/video/[video_NAME]
python3 main.py -m cnn -t video -i dataset/video/[video_NAME]

For using webcam:

python3 main.py -m ssd -t video
python3 main.py -m hog -t video
python3 main.py -m haar -t video
python3 main.py -m cnn -t video