Image Super-Resolution using CNN with Keras in Python

Image Super-Resolution using CNN with Keras in Python

Image Super-Resolution using CNN with Keras in Python

Enhance/Upsample Images with Convolutional Neural Network for Computer Vision With TensorFlow on Google Colab : Hands-on

Language: english

Note: 0/5 (0 notes) 708 students

Instructor(s): Karthik K

Last update: None

What you’ll learn

  • Understand the fundamentals of Efficient Sub-pixel Convolutional Neural Network (CNN)
  • Build and train a the super-resolution model using Keras with Tensorflow as a backend using Google Colab
  • Assess the performance of trained model
  • Learn to use the trained model to predict the high-resolution image of a new set of image data



  • Basic knowledge of Python Programming



Welcome to the “Image Super-Resolution using CNN with Keras in Python” course. In this project, you will learn how to create a Convolutional Neural Network (CNN) in Keras with a TensorFlow backend from scratch, and you will learn to train CNNs to enhance the quality of images significantly. Our neural network will create high-resolution images from low-resolution images. Please note that you don’t need a high-powered workstation to learn this course. We will be carrying out the entire project in the Google Colab environment, which is free. You only need an internet connection and a free Gmail account to complete this course. This is a practical course, we will focus on Python programming, and you will understand every part of the program very well. By the end of this course, you will be able to build and train the deep learning model using your image dataset. After that, you will also be able to use the model to predict high-resolution images on new images and visualise them. This image super-resolution course is practical and directly applicable to many industries. You can add this project to your portfolio of projects which is essential for your following job interview. This course is designed most straightforwardly to utilise your time wisely.

Happy learning.


Who this course is for

  • Beginners starting out to the field of Deep Learning
  • Industry professionals and aspiring data scientists
  • People who want to know how to write their image super-resolution code


Course content

  • Fundamentals
    • Introduction
    • Artificial Intelligence
    • Machine Learning
    • Deep Learning
    • What is Image Super-Resolution?
    • How Image Super-Resolution is done?
    • Efficient Sub-pixel Convolutional Neural Network (ESPCN)
  • Building, Evaluating and Predicting Super-Resolution Model
    • Download Dataset
    • What is inside data folder?
    • Super-Resolution Python Code
    • What is the .h5 file?
    • What is inside test folder?
    • What is inside prediction folder?
    • Enabling GPU in Google Colab
    • Is GPU connected to Colab notebook?
    • Connect Google Colab with Google Drive
    • Import Python Libraries
    • Creating Training Data Generator
    • Creating Validation Data Generator
    • Normalize the Pixels for Training and Validation Images
    • Visualize Sample Images
    • Process the Input Images and Visualize it
    • Build a CNN Model Architecture
    • Define Utility Functions to Monitor our Results
    • Peak Signal to Noise Ratio (PSNR)
    • Dataset of Test Image Paths
    • Define Callbacks
    • Visualize Model Architecture
    • Model Compilation
    • Training the Model
    • Model Testing / Evaluation
    • Prediction


Image Super-Resolution using CNN with Keras in Python

Google Project Management [Coursera with Google]

Time remaining or 291 enrolls left


Don’t miss any coupons by joining our Telegram group 

Udemy Coupon Code 100% off | Udemy Free Course | Udemy offer | Course with certificate