ABSTRACT
Image stitching, a widely employed technique in computer vision, entails merging multiple images to create a cohesive panoramic image. The feature-based method stands out as a prevalent approach in image stitching, involving the identification and matching of features across diverse images. This chapter aims to furnish a thorough overview of various feature-based methods, conducting a comparative analysis among them. The discussion delves into the algorithms associated with feature-detection methods, including SIFT, SURF, and FAST, elucidating the strengths and weaknesses of each algorithm. Additionally, guidance is provided on selecting the most suitable method based on the specific application and imaging conditions. The chapter also explores recent advancements, such as the utilization of deep learning algorithms such as convolutional neural networks, to extract intricate features from images, contributing to improved performance in the image stitching process. Furthermore, it presents a comprehensive introduction to feature-based methods, catering to researchers and practitioners in computer vision, aiding them in identifying potential research directions and choosing the optimal algorithm for their particular tasks.
