MingleNet: A Novel Dual Stacking Approach for Medical Image Segmentation

Medical image segmentation is important for disease diagnosis and treatment planning.
Ensemble learning, which combines multiple models or predictions, can improve accuracy and performance in medical image segmentation. We propose MingleNet, which uses multiple layers of ensemble learning.
MingleNet uses double-stacking of models, such as DoubleU-Net, DeepLabv3+, U-Net, and DeepLab, to produce masks.

MingleNet Architecture

Dataset

• Kvasir SEG = (https://datasets.simula.no/kvasir-seg/)
• CVC-ClinicDB = (https://datasetninja.com/cvc-612)
• CVC-ColonDB

Installations

Install all the required libraries using the command:

pip install -r requirements.txt

Note that other versions of the libraries may also work. This setup was tested with an RTX 3080 TI GPU and 32GB of RAM.

How to run

This Project can be run using the ```Notebook.ipynb``` file.

Sample Predictions

(CVC-ClinicDB)