AN EFFICIENT WAY FOR SHAPE RECOGNITION BY USING SPEEDED-UP ROBUST FEATURE (SURF)

In this paper, we present a novel scale-and rotation-invariant interest point detector and descriptor, coined CONTOUR SURF (Speeded Up Robust Features). It approximates or even outperforms previously proposed schemes with respect to repeatability, distinctiveness, and robustness, yet can be computed and compared much faster. The paper presents experimental results on a standard evaluation set, as well as on imagery obtained in the context of a real-life object recognition application. Both show SURF's strong performance. The calculation has properties of picture scaling-, interpretation, and revolution invariants. Contour-SURF feature extraction and matching technique all along with a matching practice are given for the contour based shape detection. The algorithm automatically extracts local features in the certain entity outline with no any limitation of definite neighborhood location. A Contour recognition experiment was done with different datasets. All images were taken as a target and resultant to the further model. The detection correctness reaches 100% for images having distinctive contour feature, and lesser for images having common shapes.