期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2009
卷号：9
期号：5
页码：17-29
出版社：International Journal of Computer Science and Network Security
摘要：Our day to day activities rely largely on the precise and rapid identification of objects in our visual environment. In the current scenario, Object recognition is one of the most actively researched areas of computer vision and pattern recognition. The domain of Object recognition hopes to achieve near human levels of recognition for tens of thousands of object categories under a broad variety of conditions. The significant challenge of object recognition is the ability of the system to recognize any member of a category of objects regardless of wide variations in visual appearance due to disparities in the form and color of the object, occlusions, geometrical transformations, changes in illumination, and potentially non-rigid deformations of the object itself. In this article, we investigate the effectiveness of four vital research techniques for object recognition in digital images. The techniques investigated in the article include: Principal Component Analysis, Support Vector Machines, Hidden Markov Model and k- Nearest Neighbors classifier. The comparison is examined in terms of recognition accuracy and false positives. The Columbia Object Image Library (COIL-20) is utilized in the investigation of the techniques.
关键词：Object Recognition; Object Model; Features; Principal Component Analysis (PCA); Support Vector Machines (SVM); Hidden Markov Model (HMM); k-Nearest Neighbors (k-NN); Columbia Object Image Library (COIL-20); Accuracy; False Positives
Loading...
