#43 · Primary category： Foundation Models
ModelsGenesis
Self-supervised pre-training framework for 3D medical images, learning transferable representations from unlabeled CT/MRI volumes for fine-tuning in segmentation and classification.
Project last updated：08/25/26
GitHub Stars
791
Forks
141
Contributors
10
License
MIT
Why we included this project
Training 3D segmentation or classification models on medical images usually stalls when you only have a handful of annotated scans. Models Genesis sidesteps that by pre-training on unlabeled CT and MRI volumes with image-restoration tasks, so the network learns anatomy on its own before you fine-tune it for your specific labels. The repo ships both Keras and PyTorch weights, which means you can warm-start an existing pipeline instead of reimplementing the method from the paper. It also holds up well as a baseline: the approach won a MICCAI 2019 Young Scientist Award and a MedIA Best Paper Award, and it remains a common reference point for self-supervised pretraining in medical imaging.
Articles for this project
No articles for this project yet.
To suggest a topic or contribute an article, contact us.
Related projects in this category
transformers
🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.
CLIP
CLIP (Contrastive Language-Image Pretraining), Predict the most relevant text snippet given an image
MiniCPM-V
A Pocket-Sized MLLM for Ultra-Efficient Image and Video Understanding on Your Phone
generative-models
Generative Models by Stability AI
unilm
Large-scale Self-supervised Pre-training Across Tasks, Languages, and Modalities