Computer Vision News - April 2022

29 Advanced Methods and Deep Learning in Computer Vision That will give you an understanding of the bigger picture, and then you can delve bit by bit into the details and explore those references in depth if you want to go even further. RD: While following this overall plan, it is also relevant to think carefully not only of the approach to be adopted but also how the input data (often in the form of many millions of images or image patches) should optimally be managed so as to most effectively train the final system. My aim in writing the rather long first chapter was to ensure first that readers were brought up to speed on legacy computer vision work; second, to open the doors to deep learning and to show how it can successfully be applied to computer vision; and third, to demonstrate that even when applied to the familiar rather basic subject of texture analysis that substantial changes are needed in the old ways of thinking. Am I right to assume that seasoned scholars might have an advantage in perusing the book? The Editors: Yes, absolutely. For any given chapter, only a handful of experts spend their entire time focused on the topic, but everyone in the field can learn something. If you’re an advanced person in the field but not necessarily an expert and want to learn more about a specific topic, you will benefit from this book. Is it possible that you learned things in the process of writing this book? MT: For sure. I was an editor, I didn’t write any of the chapters, but I spent a lot of time reading very carefully, providing feedback to authors, asking questions, and thinking through things myself to make sure I was understanding. Any author or editor learns a lot in the process, and I certainly did. RD: As an editor, I found it stimulating finding how best to guide eminent authors Roy Davies Matthew Turk