KonfAI: A Modular and Fully Configurable Framework for Deep Learning in Medical Imaging

2025-08-13

AI Summary

Poor reproducibility, high configuration coupling, and inadequate support for complex training paradigms hinder deep learning experiments in medical imaging. To address these challenges, we propose MedDL—a modular, declarative, and fully configuration-driven PyTorch framework. Its core contribution is a YAML-based, end-to-end declarative configuration system covering training, inference, and evaluation, which natively supports advanced paradigms including patch-based learning, test-time augmentation, model ensembling, intermediate feature extraction, and GAN training. By decoupling model architectures, loss functions, data transformations, and workflow logic, MedDL enables flexible extension and exact experimental reproduction without code modification. We validate MedDL across segmentation, registration, and synthesis tasks, where it has underpinned multiple state-of-the-art results in international competitions. The framework is open-sourced.

Technology Category

• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

KonfAI is a modular, extensible, and fully configurable deep learning framework specifically designed for medical imaging tasks. It enables users to define complete training, inference, and evaluation workflows through structured YAML configuration files, without modifying the underlying code. This declarative approach enhances reproducibility, transparency, and experimental traceability while reducing development time. Beyond the capabilities of standard pipelines, KonfAI provides native abstractions for advanced strategies including patch-based learning, test-time augmentation, model ensembling, and direct access to intermediate feature representations for deep supervision. It also supports complex multi-model training setups such as generative adversarial architectures. Thanks to its modular and extensible architecture, KonfAI can easily accommodate custom models, loss functions, and data processing components. The framework has been successfully applied to segmentation, registration, and image synthesis tasks, and has contributed to top-ranking results in several international medical imaging challenges. KonfAI is open source and available at href{https://github.com/vboussot/KonfAI}{https://github.com/vboussot/KonfAI}.

Problem

• Enables configurable deep learning workflows for medical imaging
• Supports advanced strategies like patch-based learning and model ensembling
• Facilitates reproducibility and reduces development time in medical tasks

Innovation

• Modular configurable framework for medical imaging
• YAML-based workflow without code modification
• Supports advanced strategies like patch-based learning