Project for CV, SJTU SE

Fork of nnUNet with Custom Logic

The segmentation in this project is conducted through the nnUNet framework, followed by the extraction of pituitary tumor features using the radiomics package. The final step involves designing the classifier using the scikit-learn package, resulting in an achieved classification accuracy of approximately 91%.

Brain tissue (WM, GM, CSF) segmentation using both multi-atlas and nnUNet approaches. This project was developed for a course titled "Medical Image Segmentation and Applications" - MISA under MAIA master program.

MTMAUNet: Multi-Task Multi-axis Attention UNet

Official repository of "Simulating Dynamic Tumor Contrast Enhancement in Breast MRI using Conditional Generative Adversarial Networks"

A repository that allows users to apply the DLICV method to their brain imaging data.

Brain Tissue Segmentation on IBSR18 Dataset

nnUNet benchmarks for The University of California San Francisco Adult Longitudinal Post-Treatment Diffuse Glioma (UCSF-ALPTDG) dataset.

Utilities to run nnUNet more easily using containers

This is a repository hosting all models detailed in the article Brain tumour segmentation with incomplete imaging data.

This is the regular nnUNet but with three new features: 1 - Training with cyclic learning rate, producing checkpoints from different convergent minima. 2 - an ensemble of the different checkpoints is used to determine uncertainty of each fold. 3 - On inference prediction is made using the lowest uncertainty prediction from 5 folds.

A Deep Learning Framework for Analysis of the Eustachian Tube and the Internal Carotid Artery

Code for implementing and comparing the Generalized Surface Loss to other loss functions

Brain tissue segmentation using several algorithms: EM, Atlas-based methods, nnunet

nnUNet benchmarks for The University of California San Francisco Brain Metastases Stereotactic Radiosurgery (UCSF-BMSR) dataset.

nnUNet paddlepaddle version

Official repository for "Pre- to Post-Contrast Breast MRI Synthesis for Enhanced Tumour Segmentation"

Playing with SimpleITK and nnU-Net to process data from the CHAOS challenge on Google Colab.

3D Slicer nnUNet integration to streamline usage for nnUNet based AI extensions.