Computer and Robot Vision
Multiresolution Gray-Scale and Rotation Invariant Texture Classification with Local Binary Patterns
IEEE Transactions on Pattern Analysis and Machine Intelligence
A new K-View algorithm for texture image classification using rotation-invariant feature
Proceedings of the 2009 ACM symposium on Applied Computing
Learning to segment images using region-based perceptual features
CVPR'04 Proceedings of the 2004 IEEE computer society conference on Computer vision and pattern recognition
| Hi-index | 0.00 |
Image texture classification plays an important role in the computer processing of images. Many methods including the Gray Level Co-Occurrence (GLCO), Local Binary Patterns and K-View algorithms for image texture classification have been proposed. There exist several variations of the K-view algorithms. Among them, the K-View algorithm using rotation-invariant features can produce higher classification accuracy; however, the algorithm is not stable. This paper presents an improved K-View Algorithm using new characteristic views selection methods. We propose three different methods for selecting characteristic views. The first method chooses characteristic views based on the interval of minimum mean gray level and maximum mean gray level of all views in the primitive set and then randomly select the same number of views in each sub-interval. The second method divides the sorted interval with equal sub-interval and then select one view randomly in each sub-interval. The third method chooses views in an equal distance starting from the minimum gray level in the sorted list. Experimental results show that the proposed characteristic views selection methods improve the performance of the K-View algorithms. This improved K-View algorithm is more robust and accurate compared with the results of the previous K-View algorithms.