Feature Extraction and Classification

Description
This repository is dedicated to processing and classifying prostate cancer images using Sparse Variational Gaussian Processes (SVGP). Future work will include the implementation of Sparse Variational Gaussian Process Classification Regression (SVGPCR).

Table of Contents

1. Introduction
2. Features
3. Installation
4. Usage
5. Repository Structure
6. Contributing

Introduction

Prostate cancer image analysis is a key area in medical imaging research. This repository aims to provide a framework for extracting features and classifying these images using advanced machine learning techniques, specifically utilizing SVGP.

Features

• Image Processing: Tools to preprocess prostate cancer images for feature extraction.
• Feature Extraction: Automated feature extraction using deep learning models. This part is an implementation of the framework https://github.com/arneschmidt/ssl_and_mil_cancer_classification 'Efficient Cancer Classification by Coupling Semi Supervised and Multiple Instance Learning' This is the implementation of the code of the paper A. Schmidt, J. Silva-Rodríguez, R. Molina and V. Naranjo, "Efficient Cancer Classification by Coupling Semi Supervised and Multiple Instance Learning," in IEEE Access, vol. 10, pp. 9763-9773, 2022, doi: 10.1109/ACCESS.2022.3143345.
• Classification: Classify images using Sparse Variational Gaussian Processes (SVGP).
• Future Work: Implementation of SVGP Classification for Crowdsourcing (SVGPCR) for enhanced performance, taking advantage to the biggest crowdsourcing dataset for prostate cancer computer vision classification: https://www.sciencedirect.com/science/article/pii/S0169260724004656?via%3Dihub.

Installation

1. Clone this repository:

   git clone https://github.com/AMorQ/Feature_Extraction_and_Classification.git

2. Navigate to the repository directory:

   cd Feature_Extraction_and_Classification

3. Install the required dependencies:

   pip install -r requirements.txt

Usage

(IN ORDER OF UTILIZATION):

• main.py: call all functions
• predata.py: create correct folder structure for keras data generator
• data.py: create data generators and call feature extraction models
• data_utils.py: helper functions of data.py
• model_conv.py: feature extraction and classification models
• SVGP_utils.py: helper functions of SVGP classification
• metrics.py: calculate metrics
• mlflow_logging: log parameters and metrics to MLFLOW

Repository Structure

Feature_Extraction_and_Classification/
├── data/                # Directory for storing input data
├── models/              # Contains pre-trained and custom models
├── scripts/             # Scripts for preprocessing and training
├── results/             # Directory for storing output results
├── requirements.txt     # Python dependencies
├── model_conv.py        # Main script for feature extraction and classification
└── README.md            # Repository documentation