teoshibin / COMP3007_CV_Face_Recognition

Face Recognition using VGGFace & Template Matching in Matlab

Face Recognition

University coursework, a re-implementation of face recognition models in Matlab

Reproducibility

1. MATLAB Toolboxes

   • deep learning toolbox
   • parallel computing toolbox
   • Converter for TF Keras Model

2. Model Params

   • MATLAB VGGFace Params - place it under Weights\VGGFace\ for it to work

Dataset

• LFW funneled | kaggle | official
• a small subset of celeb dataset (100 subjects 1344 test images)

Folder Structure

.
└── root/
    ├── Common/
    │   ├── DatabaseFunctions/
    │   ├── GeneralFunctions/
    │   ├── lfw-details/
    │   └── testLabel.mat
    ├── FaceDatabase/
    │   ├── lfw-funneled/
    │   ├── ...
    │   ├── Train/
    │   │   ├── example_subject_1/
    │   │   └── example_subject_2/
    │   └── Test/
    ├── Models/
    ├── Weights/
    ├── Screenshot/
    ├── Evaluation.m
    ├── EvaluationAll.m
    ├── GenerateROC.mlx
    └── README.md

Common includes all auxiliary functions for dataset manipulation and dlmodel functions
lfw-details extracted text files from LFW Dataset
testLabel.mat true label values for prepared train and test set
FaceDatabase image dataset dump
Models contain subfolders of essential model functions
Weights contain subfolders of essential model pretrained parameters
Screenshot screenshots
Evaluation.m main script
EvaluationAll.m extra main script
GenerateROC.mlx ROC plot live script

License

Licensed under the MIT License.