基于纹理和形状特征的生物图像分类.pdfVIP

基于纹理和形状特征的生物图像分类 摘要 随着生物图像数据的增加，研究人员急切 要使用计算机技术来对生物图像进行智 能化的分类。本文的研究内容即为基于纹理和形状特征的生物图像识别，主要工作如 下： 首先，本文系统地描述了生物图像分类的背景及其意义，对目前的研究情况及在相 关领域内的成果进行了简单说明，并对生物图像特征提取的基本方法以及图像模式识 别做了详尽的阐述。 其次，针对图像的形状特征，本文介绍了边缘检测的五种方法，分析其优缺点，并 最终选定canny 算子作为边缘检测方法。针对图像的纹理特征，本文分析了基于Gabor 滤波器的纹理提取的方法，并采用Gabor 滤波器作为纹理提取方法。提取完成后，以 图像的形状和纹理特征为基础，构成一个特征集。利用 人工神经网络模型进行分类， 并提出了基于人工神经网络的识别方案。 最后，本文对实验过程中出现的问题进行了反思，并对图像分类技术进行了展望。 关键词：边缘检测，Canny 算子，特征提取，Gabor 滤波，数字图像处理 Abstract As the biological image data increases, the researchers need to use computer technology to classify the biological images intelligently.. The research content of this paper is based on texture and shape feature of biometric image recognition, the main work is as follows: First of all, this paper systematically describes the background and significance of biological image classification, the present research situation and in related fields in the results of a simple description, and gave a detailed description to image features are extracted from the basic method and image pattern recognition. Secondly, for the shape feature of the image, five methods of edge detection are introduced, the advantages and disadvantages of the image are analyzed, and the Canny operator is chosen as the edge detection method.. In view of the texture feature of the image, this paper analyzes the texture extraction method based on Gabor filter, and adopts Gabor filter as texture extraction method. After the extraction, the shape and texture features of the image are the basis, and constitute a feature set.. The ANN model is classified and the recognition scheme based on ANN is proposed. At last, this paper reflects the problems in the experiment, and the image classification technology is prospected. Key Words: Edge detection, Canny operator, Feature extraction, Gabor filter,Digital image proce