Computer Vision News - April 2022

33 Advanced Methods and Deep Learning in Computer Vision important, but collections like this book or the basics of the introductory textbooks are still vital and relevant for our field. RD: It is interesting that the field changed so radically around 2012, and this raises the question of whether advances will now once again proceed at a much steadier pace. In fact, some chapters in this book seem to indicate not: in particular, chapters 12, 13 and 15 raise the relevance of cognitive theories, self-aware systems and ‘adversarial’ methods. Readers would do well to keep such new directions in mind and maybe wonder whether further developments in these directions will take us into totally new worlds that reflect the parts of our brains that are not so narrowly focussed on vision: after all, vision is only one aspect of our cognitive evolution. school level, and its resolution must not simply be left to chance. Thank you for a fascinating interview, Matthew and Roy. Do you have a final message for our readers and the wider community? MT: This field is moving very quickly. When you look back over the last decade, there’s been a massive amount of progress and a significant change in focus. Still, these fundamental pedagogical textbooks, like Roy’s book and Richard Szeliski’s book , are critical to the field. Having that broad background of what has built this field over the years is still important. Our book and other books like it, which go into a handful of topics in some depth and breadth, will be relevant for a long time to come. The field may be moving quickly, but that doesn’t mean people should focus solely on the latest conference papers. Those are Rama Chellappa at BMVC 2019 in Newcastle, UK. He and his co-authors contributed Chapter 7 of the book.