ALL UPDATES ARE COMING SOON......
The Maskvariate introduces a novel approach for contrastive loss along with it combines contrastive loss with the Dice loss to improve spatial accuracy and feature representation in medical image segmentation, particularly in scenarios with limited annotations, leading to increased precision in the analysis of medical images.
- Python 3.12.0
- TensorFlow 2.12.0
Clone the repository and install the dependencies:
git clone https://github.com/SUNNY11286/Maskvariate.git
** install all compatible version of other libraries with given prereuisites
We utilized high-quality multi-modal MRI scans from the BraTs 2016 and BraTs 2017 datasets, including 274 and 285 patients, respectively. Emphasizing the FLAIR modality, our dataset encompasses diverse tumor shapes, sizes, and locations. Additionally, to assess the generalizability of our approach, we integrated the AbdomenCT-1K dataset, comprising over a thousand abdominal CT scans, into our study.
Charulkumar Chodvadiya, Navyansh Mahla, Dr Kshitij Jadhav