Official implementation of DLCE-MIL: Depth-aware Local Context Enhancement for Patient-Level IVCM Classification.
This repository contains the code for our MICCAI 2026 paper on patient-level fungal keratitis subtype classification using in vivo confocal microscopy (IVCM) images.
DLCE-MIL is a multiple instance learning framework for patient-level IVCM classification. It combines frozen DINOv2 visual features, local instance context enhancement, and depth-aware feature fusion.
The code and instructions are being organized and will be released soon.
Due to privacy and ethical restrictions, the clinical IVCM dataset used in this study cannot be publicly released.
If you find this work useful, please cite our paper:
@inproceedings{men2026dlcemil,
title={DLCE-MIL: Depth-aware Local Context Enhancement for Patient-Level IVCM Classification},
author={Men, Xiaoyan and others},
booktitle={MICCAI},
year={2026}
}