Performance evaluation of feature detection in using subsampled images for image stitching

Image stitching is a technique being used to stitch multiple images together to form a stitched image with higher resolution and larger field of view. It has been adopted in many research areas such as computer vision and medical applications. The first step in stitching is to first determine the overlap region or similarity. This can be done using feature extraction, follow by comparing the feature points in the images. Once the similarity is determined, the images are transformed into the same coordinate system that agreed by each other. Finally, the images can be pasted together. In this paper, the performance of various type of feature detector in extracting feature points from the subsampled version of the original images will be evaluated. The advantage of using the subsampled images is to reduce the memory requirement, and processing time.