Art Style Transfer with GANs

Deep Learning, TensorFlow, keras, convolutional neural networks, neural art style transfer, Google Deep Dream
Rajeev Ratan
6 weeks · 7-10 hours per week
In this liveProject, you’ll explore the capabilities of an AI algorithm to create beautiful art. Following research laid out in a groundbreaking paper, you plan to create an algorithm that can take the aesthetic style of one image and apply it another. This fun tool can make photos look like paintings, and also augment image datasets for training other AI. To create this AI, you explore the latent space of a deep neural network, and manipulate its values to see how it affects an input image. Your challenges will include training an image classifier, manipulating your filters to produce dreamlike images, and creating AI-generated images that look like human art.

project author

Rajeev Ratan
Rajeev Ratan is a data scientist, and computer vision consultant and researcher. He has spent the last five years working at several computer vision startups, and has created several popular online courses on OpenCV and deep learning convolutional neural networks. He holds an MSc in Artificial Intelligence from the University of Edinburgh, and has published research on using data-driven methods for Probabilistic Stochastic Modeling for Public Transport.

Prerequisites

This liveProject is for intermediate Python programmers looking to enhance their data science skills with image manipulation techniques. To begin this liveProject, you will need to be familiar with:

TOOLS
  • Basics of TensorFlow
  • Basics of Keras
  • Basics of scikit-learn
  • Basics of Jupyter Notebook

TECHNIQUES
  • Basics of Computer Vision
  • Basics of Deep Learning
  • Basis of Linear Algebra

you will learn

In this liveProject, you'll utilize popular Python deep learning tools to build artistically-inclined algorithms. These popular tools and techniques are easily applied to other deep learning tasks common in industry.

  • Building convolutional neural networks with TensorFlow and Keras
  • Analyzing your model’s performance with scikit-learn
  • Visualizing filter and class maximizations with keras-vis
  • Mathematical operations on images
  • Implementing loss functions
  • Transfer learning
  • Model manipulation

features

Self-paced
You choose the schedule and decide how much time to invest as you build your project.
Project roadmap
Each project is divided into several achievable steps.
Peer support
Chat with other participants within the liveProject platform.
Compare with others
For each step, compare your deliverable to the solutions by the author and other participants.
Book and video resources
Excerpts from Manning books and videos are included, as well as references to other resources.

project outline

Introduction

Prerequisites Test

Introduction Video

Get Started

1. Training a Simple Image Classifier using Convolutional Neural Networks

1.1 Training a Simple Image Classifier using Convolutional Neural Networks

But What is a Neural Network?

Fundamentals of Machine Learning

1.2 Submit Your Work

2. Understanding what Convolutional Neural Networks Learn

2.1 Understanding what Convolutional Neural Networks Learn

Visualizing What Convnets Learn

2.2 Submit Your Work

3. Transfer Learning & Feature Map Visualization

3.1 Transfer Learning

3.2 Feature Map and Filter Visualization of a Pretrained Model

Using a Pretrained Convnet

Transfer Learning

3.3 Submit Your Work

4. Visualizing Filter Maximizations, Grad-CAM, and Class Maximization

4.1 Visualizing Filter Maximizations

4.2 Grad-CAM Visualizations

4.3 Class Maximization

How Convolutional Neural Networks See the World

4.4 Submit Your Work

5. Implementing Google’s Deep Dream

5.1 Implementing Google’s Deep Dream

Deep Dream

5.2 Submit Your Work

6. Implementing Neural Style Transfer

6.1 Neural Style Transfer Loss Functions

6.2 Neural Style Transfer Algorithm

Neural Style Transfer

6.3 Submit Your Work

Summary

Project Conlcusions

FAQs

placing your order...

Don't refresh or navigate away from the page.
Manning Early Access Program (MEAP) In MEAP, you get immediate access to a liveProject under development, so you can participate while it is created, tested, and improved. Get started today, and pick up right where you've left off whenever we update the project!
liveProject $30.00 $50.00 self-paced learning
Art Style Transfer with GANs (liveProject) added to cart
continue shopping
go to cart

Prices displayed in rupees will be charged in USD when you check out.