Trained Generative and Classification Models from "Generating Synthetic Past and Future States of Knee Osteoarthritis Radiographs Using Cycle-Consistent Generative Adversarial Neural Networks"

Published: 17 February 2025| Version 2 | DOI: 10.17632/rhs88fdrwg.2
Contributors:
Fabi Prezja, Leevi Annala, Sampsa Kiiskinen, Suvi Lahtinen, Timo Ojala, Paavo Nieminen

Description

This repository encompasses the fully trained Cycle-Consistent Generative Adversarial Neural Networks and the EfficientNetV2 Classification model from the paper "Generating Synthetic Past and Future States of Knee Osteoarthritis Radiographs Using Cycle-Consistent Generative Adversarial Neural Networks", published in Computers in Biology and Medicine (Elsevier) https://www.sciencedirect.com/science/article/pii/S0010482525001350. To load these models effectively, users must have TensorFlow/Keras installed and be running Python version 3.7 or higher.

Files

Steps to reproduce

Please see: https://www.sciencedirect.com/science/article/pii/S0010482525001350

Institutions

Jyvaskylan Yliopisto

Categories

Osteoarthritis, Synthetic Image, Deep Learning, Generative Adversarial Network

Licence