Skip to content

Latest commit

 

History

History
26 lines (18 loc) · 923 Bytes

README.md

File metadata and controls

26 lines (18 loc) · 923 Bytes
Raw

Keras-UNet

Coursera Deep Learning Specialization Course 4: Convolutional Neural Network

Week 3 Assignment 2

This repository provides the answer for coursera deep learning specialization course 4, week 3 assignment 2: Image Segmentation with U-Net. Dataset used here are provided in the course. You can however, use the code for training your own custom dataset, or modify it for other usage.

U-Net Architecture

Tensorflow Keras implementation of U-Net

To Run

  • Clone this repository
  • Run Image_segmentation_Unet_v2.ipynb on Google Colab (recommended)
  • Switch to GPU runtime
  • Run each cell and make sure all tests pass :)

Custom Dataset

  • Prepare own images and masks
  • In data folder:
    • Replace images in CameraMask with custom images' masks
    • Replace images in CameraRGP with custom images
    • Make sure the images and its corresponding masks have the same name