[NEWS!]This paper has been accpeted by CVPR 2023! The basic code on PyTorch has been opened!
[NOTE!!]The code will be gradually and continuously opened!
Learning inter-image similarity is crucial for 3D medical images self-supervised pre-training, due to their sharing of numerous same semantic regions. However, the lack of the semantic prior in metrics and the semantic-independent variation in 3D medical images make it challenging to get a reliable measurement for the inter-image similarity, hindering the learning of consistent representation for same semantics. We investigate the challenging problem of this task, i.e., learning a consistent representation between images for a clustering effect of same semantic features. Our Geometric Visual Similarity Learning embeds the prior of topological invariance into the measurement of the inter-image similarity for consistent representation of semantic regions.
This repository provides the official PyTorch implementation of GVSL in the following papers:
Geometric Visual Similarity Learning in 3D Medical Image Self-supervised Pre-training
Yuting He, Guanyu Yang*, Rongjun Ge, Yang Chen, Jean-Louis Coatrieux, Boyu Wang, Shuo Li
Southeast University
IEEE/CVF Conference on Computer Vision and Pattern Recognition 2023
We are gradually opening up pre-trained parameters on more datasets, please focus on our Pre-Trained Model Zoo page.
If you use this code or use our pre-trained weights for your research, please cite our papers:
@inproceedings{he2023geometric,
title={Geometric Visual Similarity Learning in 3D Medical Image Self-supervised Pre-training},
author={He, Yuting and Yang, Guanyu and Ge, Rongjun and Chen, Yang and Coatrieux, Jean-Louis and Wang, Boyu and Li, Shuo},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
pages={9538--9547},
year={2023}
}
This research was supported by the Intergovernmental Cooperation Project of the National Key Research and Development Program of China(2022YFE0116700), CAAI-Huawei MindSpore Open Fund and Scientific Research Foundation of Graduate School of Southeast University(YBPY2139). We thank the Big Data Computing Center of Southeast University for providing the facility support.