Conclusion

Abstract

The application of deep learning algorithms, especially CNNs, to computer vision problems have seen a rapid progress. This has led to highly robust, efficient, and flexible vision systems. This book aimed to introduce different aspects of CNNs in computer vision problems. The first part of this book (Chapter 1 and Chapter 2) introduced computer vision and machine learning subjects, and reviewed the traditional feature representation and classification methods. We then briefly covered two generic categories of deep neural networks, namely the feed-forward and the feed-back networks, their respective computational mechanisms and historical background in Chapter 3. Chapter 4 provided a broad survey of the recent advances in CNNs, including state-of-the-art layers, weight initialization techniques, regularization approaches, and several loss functions. Chapter 5 reviewed popular gradient-based learning algorithms followed by gradient-based optimization methodologies. Chapter 6 introduced the most popular CNN architectures which were mainly developed for object detection and classification tasks. A wide range of CNN applications in computer vision tasks, including image classification, object detection, object tracking, pose estimation, action recognition, and scene labeling have been discussed in Chapter 7. Finally, several widely used deep learning libraries have been presented in Chapter 8 to help the readers to understand the main features of these frameworks.