Research on wide baseline stereo matching based on PCA-SIFT

SIFT feature matching algorithm, which has good robustness to image rotation and zoom scale, is a certain degree of stability on the angle changes and affine transformation. It also has the more stable match capacity on images which are taken from any point of view. It has been widely used in the field of wide baseline stereo matching. But the biggest problem is that extracts too many feature points, template too big and takes up more the memory. The higher dimension is also to make the matching speed down. To solve this problem, it constitutes PCA-SIFT which combined with principle component analysis (PCA) and SIFT. Compared with SIFT algorithm, the descriptors provide significant benefits in storage space and matching speed. It can reduce feature points, lower dimension of feature vector and improve matching speed. Experiments demonstrate that PCA-SIFT descriptors are more distinctive, more robust to image deformations, and more compact than standard SIFT.