DETECÇÃO E RECONHECIMENTO DE OBJETOS EM IMAGENS UTILIZANDO ALGORITMOS DE EXTRAÇÃO DE PONTOS CHAVE

Abstract

This paper presents a comparative study of descriptors keypoint algorithms doing a combination of detectors and descriptors. We combined the SIFT, SURF and Harris detector algorithms with the SIFT and SURF descriptor algorithms. Three experiments with images with different characteristics were performed, in two of these experiments were used images of real objects, in a third experiment, some degradation in an image such as scale reduction, rotation, blurring, darkening and addition of noise have been performed, to analyze the behavior of the algorithms. Experiments were also performed with all degradations together. To analyze the proposed combination, we extracted processing time and the number of inliers. This work contributes to applications that require using keypoint descriptors algorithms in detection and recognition of objects in images.

Keywords

• Object Recognition
• Local Features
• Keypoint Detector
• Keypoint Descriptors