Abstract
Image matching is an important problem in computer vision and many technologies based on local descriptors have been developed. In this paper, we propose a novel local features descriptor based on an adaptive neighborhood and embedding Zernike moments. Instead of a fixed-size neighborhood, a size changeable neighborhood is introduced to detect the key-points and describe the features in the frame of Gaussian scale space. The radius is determined by the scale parameter of the key-point and the dominant direction is computed based on skew distribution fitting instead of the traditional eight-direction statistics. Then a 72-dimensional features vector based on a 3 × 3 grid is presented. A 19-dimensional vector consists of Zernike moments is applied to achieve better rotation invariance and finally contributes to a 91-dimensional descriptor. The accuracy and efficiency of proposed descriptor for image matching are verified by several numerical experiments.
Highlights
Image matching is one of the core tasks in computer vision and it is the basis of many subsequent applications, it often refers to get the mapping between two different images by analyzing the pixel values, structures, textures, etc
Local binary patterns (LBP) is considered among the most computationally efficient high-performance texture features [23], [33], [34], [40].It is often sensitive to image noise and unable to capture macrostructure information
The proposed algorithm is compared with seven commonly used and well recognized algorithms, namely, Harris, Scale-invariant feature transform (SIFT), BRISK, speed up robust features (SURF), FREAK, LBP and local ternary patterns (LTP) on three measurements, correct matching number num, accuracy acc., cost time per feature τ
Summary
Image matching is one of the core tasks in computer vision and it is the basis of many subsequent applications, it often refers to get the mapping between two different images by analyzing the pixel values, structures, textures, etc. A retinainspired key-point descriptor is proposed to enhance the performance and a cascade of binary strings is computed by efficiently comparing image intensities over a retinal sampling patter [1] Besides of these scale invariant matching methods, several attempts have been made to create local image descriptors invariant to affine transformations [28], [29], [37]. Local binary patterns (LBP) is considered among the most computationally efficient high-performance texture features [23], [33], [34], [40].It is often sensitive to image noise and unable to capture macrostructure information These techniques first compute the difference between a pixel with the neighborhood, and apply a binary pattern to compute the represent values.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
