Mixup-VQ-VAE: A Novel Image Augmentation Technique for Clinical Machine Learning Applications

Overview

In many applications of clinical machine learning, available datasets may not be sufficiently large enough to train generalizable supervised machine learning models. To improve the generalizability of supervised machine learning models in clinical machine learning applications, Mixup-VQ-VAE, a novel data augmentation mechanism that recombines samples in the latent space of a Vector-Quantized Variational Autoencoder (VQ-VAE) using Mixup, a convex interpolation technique, is proposed. Performance of deep neural networks trained on Chest X-ray medical imagery, augmented with Mixup-VQ-VAE and other computer vision data augmentation algorithms, are compared. It was observed that Mixup-VQ-VAE improves the test accuracy of ResNet-50 on this task by 18% over the same model trained on non-augmented data, highlighting the efficacy of this augmentation mechanism for clinical machine learning applications.

Competition history

JSHS 2022 Category not listed

Resources

National JSHS abstract book

Related projects

ISEF · 2025

Convolutional Clarity: Accurate, Reliable, Explainable CNNs by Evaluating the Impact of Data Augmentation on Chest X-Ray Pneumonia Classification

ISEF · 2022

COVID-19 Detection on Chest X-Ray Using an Enhanced Neural Network Model: Impact of Network Architecture Complexity, Data Augmentation, and Transfer Learning on Model

JSHS · 2024

Iterative Misclassification Error Training (IMET): An Optimized Neural Network Training Technique for Medical Image Classification

ISEF · 2024

Iterative Misclassification Error Training (IMET): An Optimized Neural Network Training Technique for Medical Image Classification

ISEF · 2025

WeCAViT: Weighted Ensemble of CNN, Attention, and a Visual Transformer - A Novel Ensemble Neural Network Architecture for Accurate Pneumonia Diagnosis From Chest X-Rays

ISEF · 2018

Mixup-VQ-VAE: A Novel Image Augmentation Technique for Clinical Machine Learning Applications

Overview

Competition history

Resources

Related projects

Convolutional Clarity: Accurate, Reliable, Explainable CNNs by Evaluating the Impact of Data Augmentation on Chest X-Ray Pneumonia Classification

COVID-19 Detection on Chest X-Ray Using an Enhanced Neural Network Model: Impact of Network Architecture Complexity, Data Augmentation, and Transfer Learning on Model

Iterative Misclassification Error Training (IMET): An Optimized Neural Network Training Technique for Medical Image Classification

Iterative Misclassification Error Training (IMET): An Optimized Neural Network Training Technique for Medical Image Classification

WeCAViT: Weighted Ensemble of CNN, Attention, and a Visual Transformer - A Novel Ensemble Neural Network Architecture for Accurate Pneumonia Diagnosis From Chest X-Rays

Improving Robustness of X-Ray Synchrotron Image Analysis Using Deep Learning and Data Augmentation

Using A Novel Semi-Supervised Machine Learning Method to Improve Image-based Lung Cancer Diagnostic Algorithms

COVID-19 Chest X-ray Images: Lung Segmentation and Diagnosis Using Neural Networks