FEATURE POINT CORRESPONDENCE BETWEEN CONSECUTIVE FRAMES BASED ON GENETIC ALGORITHM

Y. Wang

Keywords

Long image sequence, feature point correspondence, α–β–γ filter,genetic algorithm

Abstract

This paper proposes a genetic algorithm-based approach to feature point correspondence between consecutive frames in a long image sequence containing outliers and occlusions. α–β–γ filter-based visual tracking is proposed to track the motion of feature points between the consecutive image frames according to the characteristics of high speed of image sampling and motion inertia. Then, the chromosome and the fitness function of the genetic algorithm are constructed based on factors such as tracking results of feature point, occlusion problem, outliers, and motion smoothness in image frames. Finally, the correspondence is obtained by using the genetic algorithm. Experimental results from a real image sequence demonstrate the effectiveness of the proposed approach for removing outliers and recovering occluded points.

Important Links:



Go Back