Centennial College
COMP 257 - Unsupervised and Reinforcement Learning - Final Group Project - Fall 2024

• About
• Timeline
• Authors

This project implements a facial recognition and identification system using the umist_cropped.mat dataset. The system leverages state-of-the-art data preprocessing, clustering, and neural network techniques to classify and identify individuals accurately.

• Explore the data
• Preprocess the data
  • Handle imbalanced data
  • Scale/Normalize
  • Reduce dimensionality (PCA, UMAP, Autoencoders)
• Cluster the data
  • KMeans
  • DBSCAN
  • Gaussian Mixture
  • Hierarchical Clustering
  • Spectral Clustering

• Apply neural network model to train and predict
  • CNN

• Tuong Nguyen Pham - @Currybon30
• Yi-Chen Hsu - @Janjanhsu
• Nguyen Ngoc Quynh Nhu - @NhuNhuNguyen
• Suaverdez Joan - @jsuaverd
• Tang Kwok Wing - @winghk00
• Yang Dongheun - @DanielY314