Learning OpenCV 4 Computer Vision with Python 3: Get to grips with tools, techniques, and algorithms for computer vision and machine learning, 3rd Edition

I spent several weeks researching OpenCV books, ordered a few, and was terribly disappointed. Then I stumbled onto this gem. It covers all the major topics under the OpenCV umbrella and does so succinctly and in a very readable manner. In fact, I read the whole book cover to cover without touching the code. This is almost impossible with most computing books, but the authors have a very approachable writing and teaching style. The only place they really slipped up was in the descriptions of SIFT, SURF, FAST, and similar algorithms. Those sections were pretty confusing and unclear. One thing I appreciated was that the authors take time to build a suite of modules, like you'd need in real life applications. More than just providing code, they provide guidance on when to make separate modules and why (generally to facilitate expansion of the program later). I was particularly interested in detecting custom objects, and had become very frustrated with the approaches in other books and in online tutorials. The authors' approach is straightforward and easy to follow, and I am currently applying it. Unlike other Packt books, I had no problem downloading the book's code, however, I did run into some difficulty finding and downloading some of the public training datasets (such as for detecting cars). Despite this, the book is still head-and-shoulders above other OpenCV resources I have attempted to use. If you are just getting started with OpenCV, do yourself a favor and get a copy of Learning OpenCV.

Learning OpenCV 4 is a strong introduction to computer vision with Python. It gives you a solid overview of the core fundamentals without overwhelming you with unnecessary complexity. The book walks through key concepts like image processing, transformations, filtering, feature detection, and basic machine learning applications in a structured, practical way. What makes it especially useful is that it balances theory and implementation. You get enough explanation to understand what’s happening under the hood, but the focus stays on actually using OpenCV with Python. It doesn’t dive too deeply into advanced mathematics or research-level detail, which makes it ideal for beginners or developers who want a working understanding before moving on to more advanced material.

Arrived in perfect condition. I Suggest this book to anyone learning open CV with python

I have taken two graduate level courses in image processing at Arizona State University which have taught me many of the concepts in this book like depth estimation, edge detectors, and convolutional filters. The book briefly describes the concepts of the algorithms and then shows the code necessary to perform them. It is a perfect supplement for an intermediate user who is aware of some image processing techniques/concepts but has never performed them with code, like myself. I highly recommend this book. "But can't I just look up tutorial online about OpenCV?" You certainly can and you will probably find what you are looking for. However, when searching online you have to know what you are looking for. With this book, the information comes to you rather than you seeking it out. There are so many algorithms that I did not know about, like the "Harris corner detection algorithm" that I would never come across if it weren't for this book.

This book is good for an overall review, but if lacks detail about values it passes to function calls and why using one over another. I can read the help for the cv2, and this does not do much than more than that.

This is a good book. It really does engage the reader in a few pages. As basics are important, I want to focus on Chapter 2 and the pros and cons of the teaching approach. I do see, that the author has made it clear that familiarity in Python is required, but that still is not a reason to leave subtle but important things out. I will state a couple of examples here because I feel we may have to be prepared for this pattern of teaching in subsequent chapters. So, in Chapter 2 the lesson does a great job of explaining 8-bit vs. 24-bit, 3 channel images, but then when talking about converting images to and from from byteArrays, the lesson fails to mention certain detail. For example: When reshaping an image, the height comes first, before the width Another example: When creating images from random bytes, a 120000 random numbers are chosen. Why? Why not 12000? There is no justification. And further down the line, the array is reshaped back to a 400 x 300 Grayscale image, or a 400 x 100 color image, there is no explanation on why these numbers were chosen. I had to figure that out by running into a few failures. Then I could see that 400 x 300 is chosen because the product of these 2 numbers is 120000, and 400 x 100 was chosen because it stores 3 channel colors and therefore it had to be 1/3rd of the size of the Grayscale image (400 x 100 x 3 = 120000). Inference: You can reshape the array back to an image of any width x height as long as the two equations above work out. Trying to reshape a 120000 byte array into a 300 x 500 will not work! This section also misses out on bringing the image.shape method that would give the user insight into what the shape (rows x cols) of an array is so that it can be used in determining how to work with this array. Maybe it's discussed further in the book, but all of the above bits of information would have definitely helped new users learn the basics better. We may know Python or numpy, but that does not mean we know how images as bytes are processed. Which is one of the reasons why we purchased this book I suppose. That said, the overall approach is engaging and good. But you have to discover and infer a few basics yourself.

extremely outdated

Ok

This is a great book to learn opencv library and image processing, detection etc. It makes me interested in computer vision. My favorite is the python codes comes with the book. They all work and I can easily use the codes for my application. so the book is definitely worthy! Thanks Joseph for this great book, I thorough enjoy it.

The book is very clear and easy to follow, even for new python users! Very satisfied. It’s a pity that it is in black and white, but that surely keeps the price more affordable. Great buy! :)

Sehr gutes Buch für jemanden der mit OpenCV arbeiten möchte. Man benötigt aber auch Grundkenntnisse in Python. Dafür gibts aber auch gute Bücher. Von mir eine Empfehlung.

Needed an authoritative volume on OpenCV and this looks the business. Not had much chance to work through the chapters yet but certainly seems to be plenty of depth there. Looking forward to exploring the themes.