Comprehensive interest points based imaging mosaic

This paper begins by reviewing the current state-of-the-art for imaging mosaic and applications. Then a new approach that uses a four-step automatic imaging mosaic, based on interest points, is proposed. The four steps are identification of interest points, finding corresponding points in the stitching images, deriving the spatial and spectral transform matrix then image mosaic and smoothing. The advantage of this approach is that it is robust to image capture conditions such as lighting, rotation, viewpoint and capture device.