Book Image

Modern Computer Vision with PyTorch

By : V Kishore Ayyadevara, Yeshwanth Reddy
Book Image

Modern Computer Vision with PyTorch

By: V Kishore Ayyadevara, Yeshwanth Reddy

Overview of this book

Deep learning is the driving force behind many recent advances in various computer vision (CV) applications. This book takes a hands-on approach to help you to solve over 50 CV problems using PyTorch1.x on real-world datasets. You’ll start by building a neural network (NN) from scratch using NumPy and PyTorch and discover best practices for tweaking its hyperparameters. You’ll then perform image classification using convolutional neural networks and transfer learning and understand how they work. As you progress, you’ll implement multiple use cases of 2D and 3D multi-object detection, segmentation, human-pose-estimation by learning about the R-CNN family, SSD, YOLO, U-Net architectures, and the Detectron2 platform. The book will also guide you in performing facial expression swapping, generating new faces, and manipulating facial expressions as you explore autoencoders and modern generative adversarial networks. You’ll learn how to combine CV with NLP techniques, such as LSTM and transformer, and RL techniques, such as Deep Q-learning, to implement OCR, image captioning, object detection, and a self-driving car agent. Finally, you'll move your NN model to production on the AWS Cloud. By the end of this book, you’ll be able to leverage modern NN architectures to solve over 50 real-world CV problems confidently.
Table of Contents (25 chapters)
Section 1 - Fundamentals of Deep Learning for Computer Vision
Section 2 - Object Classification and Detection
Section 3 - Image Manipulation
Section 4 - Combining Computer Vision with Other Techniques

Implementing deep Q-learning with the fixed targets model

In the previous section, we learned how to leverage deep Q-learning to solve the CartPole environment in Gym. In this section, we will work on a more complicated game of Pong and understand how deep Q-learning, alongside the fixed targets model, can solve the game. While working on this use case, you will also learn how to leverage a CNN-based model (in place of the vanilla neural network we used in the previous section) to solve the problem.

The objective of this use case is to build an agent that can play against a computer (a pre-trained, non-learning agent) and beat it in a game of Pong, where the agent is expected to achieve a score of 21 points.

The strategy that we will adopt to solve the problem of creating a successful agent for the game of Pong is as follows:

Crop the irrelevant portion of the image in order to fetch the current frame (state):

Note that, in the preceding image, we have taken the original image and cropped...