COMP 6211: Biometrics coursework

Project of biometric classification system, which is able to recognize people by their body shape. Four feature extraction techniques are used. Project was build with Python 3 and OpenCV 4.1.1.

Structure

• /datasets contains original image datasets with labeled images
• /preprocessed contains generated images from pre-proccess stage
• /templates contains image templates for template matching
• /plots contains generated plots

Installation

Use package manager pip to install dependencies

pip3 install -r requirements.txt

Usage

Pre-processing

Run preprocess.py script:

python3 preprocess.py

This script will preprocess images from /datasets and save them to /preprocessed.

Three different image types will be created:

1. body - grayscale bodyimages
2. masked_body - masked images of body shape
3. masked_head - masked images of head shape

Classification

Run model.py script for classification:

python3 model.py

Classification will print accuracies for individual classifiers and combined accuracy. Three plots are created.

Inter/Intra class variation (combined results)

Equal Error Rate (combined results)

Cummulative Match Curve (all classifiers)