You're reading from Modern Computer Vision with PyTorch A practical roadmap from deep learning fundamentals to advanced applications and Generative AI

Product type Paperback

Published in Jun 2024

Publisher Packt

ISBN-13 9781803231334

Length 746 pages

Edition 2nd Edition

Languages

Python

Tools

PyTorch

Concepts

Computer Vision

Authors (2):

V Kishore Ayyadevara

Yeshwanth Reddy

View More author details

Table of Contents (26) Chapters

Preface

1. Section 1: Fundamentals of Deep Learning for Computer Vision

2. Artificial Neural Network Fundamentals FREE CHAPTER

3. PyTorch Fundamentals

4. Building a Deep Neural Network with PyTorch

5. Section 2: Object Classification and Detection

6. Introducing Convolutional Neural Networks

7. Transfer Learning for Image Classification

8. Practical Aspects of Image Classification

9. Basics of Object Detection

10. Advanced Object Detection

11. Image Segmentation

12. Applications of Object Detection and Segmentation

13. Section 3: Image Manipulation

14. Autoencoders and Image Manipulation

15. Image Generation Using GANs

16. Advanced GANs to Manipulate Images

17. Section 4: Combining Computer Vision with Other Techniques

18. Combining Computer Vision and Reinforcement Learning

19. Combining Computer Vision and NLP Techniques

20. Foundation Models in Computer Vision

21. Applications of Stable Diffusion

22. Moving a Model to Production

23. Other Books You May Enjoy

24. Index

Appendix

Understanding autoencoders

So far, in previous chapters, we have learned about classifying images by training a model based on the input image and its corresponding label. Now let’s imagine a scenario where we need to cluster images based on their similarity and with the constraint of not having their corresponding labels. Autoencoders come in handy for identifying and grouping similar images.

How autoencoders work

An autoencoder takes an image as input, stores it in a lower dimension, and tries to reproduce the same image as output, hence the term auto (which, in short, means being able to reproduce the input). However, if we just reproduce the input in the output, we would not need a network, but a simple multiplication of the input by 1 would do. The differentiating aspect of an autoencoder from the typical neural network architectures we have learned about so far is that it encodes the information present in an image in a lower dimension and then reproduces the...