Performance Evaluation of Linear and Nonlinear Scale Descriptor Based Algorithms

Feature extraction and object matching is one of the important and significant task in many computer vision and robotic applications. Finding the corresponding points between the two images is very crucial in image analysis. The KAZE and SURF algorithms are found to perform well on feature matching. This paper investigates the performance evaluation of linear scale and nonlinear scale based algorithms namely KAZE and SURF respectively. The experimental results show that KAZE is more efficient than SURF but takes more computational resources.