This paper presents a real-time keypoint matching algorithm using a local descriptor derived by Zernike moments. From an input image, we find a set of key-points by using an existing corner detection algorithm. At each keypoint we extract a fixed size image patch and compute a local descriptor derived by Zernike moments. The proposed local descriptor is invariant to rotation and illumination changes. In order to speed up the computation of Zernike moments, we compute the Zernike basis functions in advance and store them in a set of lookup tables. The matching is performed with an Approximate Nearest Neighbor (ANN) method and refined by a RANSAC algorithm. In the experiments we confirmed that videos of frame size 320x240 with the scale, rotation, illumination and even 3D viewpoint changes are processed at 25~30Hz using the proposed method. Unlike existing keypoint matching algorithms, our approach also works in real-time for registering a reference image.
Local Descriptor by Zernike Moments for Real-Time Keypoint Matching
2008 Congress on Image and Signal Processing ; 2 ; 781-785
2008-05-01
1184199 byte
Conference paper
Electronic Resource
English
Keypoint descriptor matching with context-based orientation estimation
British Library Online Contents | 2014
|Terrain Matching by Fusing HOG With Zernike Moments
IEEE | 2020
|Efficient Terrain Matching With 3-D Zernike Moments
IEEE | 2019
|Improvement and Invariance Analysis of Zernike Moments Using As a Region-Based Shape Descriptor
British Library Online Contents | 2002
|Neural Network Image Recognition by Zernike and Pseudo-Zernike Moments
British Library Online Contents | 2003
|