Skip to content

Latest commit

 

History

History
28 lines (23 loc) · 2.05 KB

README.md

File metadata and controls

28 lines (23 loc) · 2.05 KB

A comparison between SegNet and U-Net

In this repository, i have tried to impliment SegNet and U-Net and tried to compare their results that we have got for segmentation.

Dataset

The dataset that I have used is a kaggle dataset for identification and segmentation of nuclei in cells. The dataset consists of 670 images and each of the image is an RGB image with dimension 512×512. Below is the figue of the type of data that we have for segmentation. Fig: Row1 consists of original images followed by row2 that consists of original labelled images(Source:Kaggle)

Segmentation by SegNet

The network architecture that i have used here has VGG-166 in downsampling. The network architecture consists of two convolutional layers of kernel (3,3) followed by a maxpool. Fig: SegNet architecture; Source: google

Segmentation by U-Net

The main difference of SegNet and U-net is that we use concatenation of layers in case of U-Net that helps the network to learn better while upsampling the data. Below is the figure of a simple U-Net architecture: Image result for U-Net architecture Fig: An architecture of a U-Net Network; Source:google

Results

Fig: Row1: Input data,Row2: SegNet output,Row3: U-Net output,Row4: Original labels