Overview / Usage

U-net architecture can be applied to various other segmentation tasks not just limited to medical images. It achieves a higher accuracy than the state of the art segmentation methods. It requires very few images and takes less time for training.

Methodology / Approach

A U-net consists of various convolutional layers. But, the pooling operators are replaced by upsampling operators which increases the resolution of the output.

Technologies Used

Tensorflow
Python