U-Net: Convolutional Networks for Biomedical Image Segmentation
This repository created for CSI 416 (Pattern Recognition Lab) projects.
Our course faculty is # Sajid Ahmed.
And Our team members are
- Mazharul Islam Leon
- Md Ifraham Iqbal
- Sadaf Meem: Github Link update: https://github.com/Iammeem02/u_net_pattern_lab
- Sharmin Sultana
- Sanjana Rahman
In this repostitory we are trying to build the U-Net algorithm. We are also trying to build U-Net in different framework like keras, pytorch, fastai. For datasources, we use kaggle api to download data from kaggle. For implementation. we use colab beacause colab is easy to use and also support free GPU. Basically, we are trying to find out how u-net algorithm is works and we hope that we already learned a lot with this state-of-the-art algorithm.
U-Net :
- U-net algorithm is very efficient with image segmentation. For bio-medical image segmentation, to get the better result using deep learning approach, some of the models are not perform well. The main reason behind that problem is low number image data. And U-Net model performs better with low datasets like these bio-medical images. U-Net model is the winner of various medical data science competition.