Invariant texture classification using a spatial filter bank in multi-resolution analysis |
| |
Affiliation: | 1. School of Computer Engineering, Iran University of Science and Technology (IUST), Narmak, 16846-13114 Tehran, Iran;2. Biomedical Engineering Department, School of Electrical Engineering, Iran University of Science and Technology (IUST), Narmak, 16846-13114 Tehran, Iran;1. Department of Computer Science and Engineering, Key Laboratory of Computer Vision and System, Ministry of Education, Tianjin University of Technology, No 319 Binshui West Road, Tianjin 300384, China;2. B-DAT and CICAEET, School of Information and Control, Nanjing University of Information Science and Technology, 219 Ningliu Road, Nanjing, Jiangsu 210044, China;3. B-DAT, School of Information and Control, Nanjing University of Information Science and Technology, Nanjing 210044, China |
| |
Abstract: | This paper proposed a new method based on spatial filter banks and discrete wavelet transform (DWT) for invariant texture classification. The method used a multi-resolution analysis method like DWT and applied the proposed filter bank on different resolutions. Then, a simple fusion of features on different resolutions was used for invariant texture analysis. A comprehensive study was done to examine the effectiveness of the proposed method. Different datasets with different properties were used in this paper such as Brodatz, Outex, and KTH-TIPS for the evaluation. Local binary pattern (LBP) methods have been one of the powerful methods in recent years for invariant texture classification. A comparative study was performed with some state-of-the-art LBP methods. This comparison indicated promising results for the proposed approach as compared with the LBP methods. |
| |
Keywords: | |
本文献已被 ScienceDirect 等数据库收录! |
|