期刊名称：Journal of Automation, Mobile Robotics & Intelligent Systems (JAMRIS)
印刷版ISSN：1897-8649
电子版ISSN：2080-2145
出版年度：2017
卷号：11
期号：1
页码：3
DOI：10.14313/JAMRIS_1-2017/1
出版社：Industrial Research Inst. for Automation and Measurements, Warsaw
摘要：The usage of real-valued, local descriptors in computer vision applica.ons is oōen constrained by their large me- mory requirements and long matching .me. Typical ap- proaches to the reduc.on of their vectors map the des- criptor space to the Hamming space in which the obtai- ned binary strings can be e.ciently stored and compa- red. In contrary to such techniques, the approach pro- posed in this paper does not require a data-driven bi- narisa.on process, but can be seen as an extension of the .oa.ng-point descriptor computa.on pipeline with a step that allows turning it into a binary descriptor. In this step, binary tests are performed on values determined for pixel blocks from the described image patch. In the paper, the proposed approach is described and applied to two popular real-valued descriptors, SIFT and SURF. The paper also contains a comparison of the approach with state-of-the-art binarisa.on techniques and popu- lar binary descriptors. The results demonstrate that the proposed representa.on for real-valued descriptors out- performs other methods on four demanding benchmark image datasets.
