The Comparison Between Two Texture Classification Approaches

Texture is one of the fundamental essential factors which constitutes the real world, and it exists in all sides of human beings life. The sense perception of texture is an important element for human beings to recognize the world. Computer is having the capacity of feeling the world around, learning its changing regular and components step by step. As an image pattern, texture naturally becomes an important problem for study.There are two purposes to study texture: one is its observation; the other is its image character. As for computer vision, texture is a visional label to segment and discern the scene and the surface of objects. At present, texture analysis includes three main questions: texture segmentation, texture classification, shape recovery from texture.In texture classification, the problem is identifying the given textured region from a given set of texture classes on the basis of pixels of a image belonging to different regions. Because the textures belong to different regions is not only related to the given pixel gray, but also have a close relationship with distribution of gray levels in a neighborhood. Texture classification is considered as a compound of two questions: feature extraction and classification process. The methods of feature extraction have statistical methods, model-based methods, signal-based processing methods, where we often use gray-level co-occurrence, frequency analysis and fractals.The text studies the formation of the features of texture images and puts forward a series of filter bank parameter selection rules , the purpose is to reduce the number of filters required for texture feature extraction and form the maximum discrimination between class, correlative, low dimension feature space, simplify classification complexity, increase texture classification speed under the premise of keeping texture classification result; we also give two common representation of filter outputs—textons and binned histograms, furthermore there is a correspondence between the two representation.