Extraction of Similar and Identical Feature Points in Mobile Images – A Survey

Author (s) : K.S. Aiswarya, N. Santhi and K. Ramar

---

Abstract :
Extraction of Image feature algorithms such as Speed Up Robust Features (SURF), Scale Invariance Featured Transformation (SIFT), Local Binary Pattern (LBP) and Principal Component Analysis - Scale Invariance Featured Transformation (PCA-SIFT) were studied and compared. The above algorithms are very robust and rugged. SIFT is highly stable algorithm which uses Difference of Gaussians for calculating the salient points. This algorithm is slow in comparison to other algorithms. The second algorithm is SURF which uses fast Hessian matrix to extract the salient points. This is the main advantage of SURF which makes it faster by triple times than SIFT. Next algorithm is PCA-SIFT which exhibits nice performance in illumination of elemental images and image rotation, also much faster than SIFT. Local Binary Pattern (LBP) the fourth algorithm uses comparison of neighborhood pixels and centre pixel in any image points. The computations used in this algorithm are simple, fast and robust in terms of feature extraction.

Keywords : Local Binary Pattern, Principal Component Analysis-Scale Invariance Featured Transformation

Conference Date : 16-09-2020

Pages : 1 - 8

Paper ID : 108491