Research Of Image Retrieval Based On Texture Features

With the information increasing rapidly worldwide, more and more attention had been paid to how to retrieve the visual information. Image retrieval based on content has become the important research hotspot in the current information retrieval field. There are three main researches orientation of content based image retrieval: image retrieval based on the image color features, the image texture features and the image shape features.This thesis begins with the image texture features, researches on how to extract the image texture features to descript the image effectively, and retrieve the image based on the features extracted. The target of retrieval of the texture feature is to find out the proper texture feature of the image in order to enhance the retrieve efficiency. Firstly, this thesis begins with the color space selection and the extraction of color features.It describes RGB,HSI,HSV,HSB,YUV color space, and elaborates image visual feature retrieval and the comparison method of image similarity degree. Secondly, we describe the definition and property of texture in detail and introduce three methods of texture analysis which including gray statistics method, structure method and frequency-domain method., particularly the feature description and feature extraction based on texture element pattern statistical method. and put forward gray concurrence feature extraction based on texture element pattern. We also talk about Gauss normalization and method of image similarity degree. Finally, we design and implement the image retrieval system and research each module's function in detail, the experimental results show that the feature description and feature extraction based on texture element pattern statistical can descript the image feature better, and it gains better results.