Improved polar complex exponential transform for robust local image description

Image description via robust local descriptors plays a vital role in a large number of image representation and matching applications. In this paper, we propose a novel distinctive local image descriptor that is based on the phase and amplitude information of Polar Complex Exponential Transform (PCET). The proposed descriptor, called IPCET (Improved PCET), is robust to the common photometric transformations (e.g., illumination, noise, JPEG compression, and blur) and geometric transformations (e.g., scaling, rotation, translation, and significant affine distortion). We perform extensive experiments to compare our IPCET descriptor with six most cutting-edge region descriptors (i.e., SIFT, Zernike Moment, GLOH, PCA-SIFT, SURF, and ORB). Experimental results demonstrate that our IPCET descriptor outperforms cutting-edge moment-based descriptors.

Automated summary

In this paper, a novel local image descriptor called IPCET is proposed, which is based on the phase and amplitude information of PCET. The IPCET descriptor is robust to both photometric and geometric transformations, and outperforms cutting-edge moment-based descriptors according to extensive experiments.