Finding Images In A Sequence Of Videos

Abstract

attern recognition techniques have been widely used in a variety of scientific disciplines including computer vision, image understanding, biology and so on. Although many methods present satisfactory performances for image analysis, they still have several weak points and which leave a large space room for further improvements and enhancements. For example, the linear discriminant analysis (LDA) algorithm is able to extract discriminative features, but the small sample size (SSS) problem makes its application scope limited. In this work, we have proposed several feature extraction and learning algorithms in order to improve the classification performance in image analysis. At the beginning we decided to study the performance of the SIFT Operator aiming to find the proper image also with changing some parameters so that we can see its influence, then, our next objective was to combine both the SIFT Operator with the Harris Corner trying to produce a more effective hybrid method, better than each one of them. For each proposed method, extensive experiments are designed in the interest of evaluating its effectiveness to find an image in a sequence of videos, then, comparing the performance of each one and the ability of minimizing the margin error.