结合角度径向变换和局部二值模式的纹理分类.docx

该【结合角度径向变换和局部二值模式的纹理分类 】是由【niuwk】上传分享，文档一共【3】页，该文档可以免费在线阅读，需要了解更多关于【结合角度径向变换和局部二值模式的纹理分类 】的内容，可以使用淘豆网的站内搜索功能，选择自己适合的文档，以下文字是截取该文章内的部分文字，如需要获得完整电子版，请下载此文档到您的设备，方便您编辑和打印。结合角度径向变换和局部二值模式的纹理分类
Abstract
Texture classification is a challenging problem in computer vision. In this paper, we propose a novel approach that combines the Angular Radial Transformation (ART) and Local Binary Patterns (LBP) for texture classification. The proposed method extracts both the global and local texture features of an image and employs a classification algorithm to classify the image into one of the pre-defined categories.
Introduction
Texture classification is an important task in computer vision since it has widespread applications in different areas such as medical imaging, satellite images, and remote sensing. Various texture features have been developed for texture classification including statistical features, frequency domain features, and model-based features. However, these features have limitations such as low discriminative power and high-dimensional feature vectors. To overcome these limitations, we propose a novel approach that combines the Angular Radial Transformation (ART) and Local Binary Patterns (LBP) for texture classification.
Angular Radial Transformation (ART)
Angular Radial Transformation (ART) is a global texture feature that captures the orientation and magnitude of texture variations in an image. It transforms an image into a polar coordinate system and then calculates the magnitude and orientation of the texture variations relative to the origin. The magnitude and orientation are then concatenated to form a feature vector.
Local Binary Patterns (LBP)
Local Binary Patterns (LBP) are a type of local texture descriptor that captures the local information about the texture variations in an image. LBP works by comparing the pixel values in a circular neighborhood around each pixel with the center pixel value. The result of the comparison is represented using binary digits, and the binary digits are then concatenated to form a feature vector. The LBP feature vector captures the local texture variations in an image.
Proposed Method
The proposed method combines ART and LBP to capture both the global and local texture features of an image. First, the image is transformed using ART, and the magnitude and orientation of the texture variations are extracted. Second, LBP is used to extract the local texture features of the image. The LBP descriptor is calculated for each pixel in the image, and the resulting feature vector captures the local texture variations in the image.
After extracting the global and local texture features, a classification algorithm is used to classify the image into one of the pre-defined categories. In this paper, we used a support vector machine (SVM) classifier to classify the image. The SVM classifier was trained on a set of labeled images to learn the mapping between the feature vectors and the corresponding categories. The trained SVM classifier was then used to classify the test images.
Experimental Results
To evaluate the performance of the proposed method, we conducted experiments on three widely used texture datasets: Brodatz, CUReT, and KTH-TIPS. We compared the proposed method with several other state-of-the-art texture classification methods including LBP, Gabor, and Local Gabor Binary Pattern (LGBP).
Table 1. Comparison of classification accuracy for different methods
| Dataset | Proposed Method | LBP | Gabor | LGBP |
|---------|----------------|-----|-------|------|
| Brodatz | % | % | % | % |
| CUReT | % | % | % | % |
| KTH-TIPS| % | % | % | % |
Table 1 summarizes the classification accuracy of the proposed method and other state-of-the-art methods on the three datasets. The results show that the proposed method outperforms the other methods in terms of classification accuracy. The proposed method achieves an accuracy of %, %, and % on the Brodatz, CUReT, and KTH-TIPS datasets, respectively.
Conclusion
In this paper, we proposed a novel approach that combines the Angular Radial Transformation (ART) and Local Binary Patterns (LBP) for texture classification. The proposed method extracts both the global and local texture features of an image and employs a classification algorithm to classify the image into one of the pre-defined categories. Experimental results show that the proposed method outperforms other state-of-the-art texture classification methods in terms of classification accuracy. The proposed method can be used in various applications such as medical imaging, satellite images, and remote sensing.