Deep Learning Project: Shoe Image Classification

This Deep Learning Project was prepared by:

1. Lang Kah Chun
2. Lee Jia Fui
3. Lee Jia Wei
4. Sun Qi Yang

Objective

The main objective of this project is to accurately classify different types of sneakers using deep learning techniques.

Data Collection

The interesting part of this project is that we self-collected the shoe image dataset using our own devices and compiled it into a folder. The dataset is available here.

Methodology

1. Hyperparameter Tuning: We implemented RayTune to assist with hyperparameter tuning using grid search.
2. Transfer Learning: Several transfer learning models were used to predict the class of sneakers accurately.

Summary

This project demonstrates the practical application of deep learning techniques, from data collection to model tuning and implementation. By leveraging RayTune for hyperparameter optimization and using transfer learning models, we achieved accurate classification of sneaker images.