Procesamiento Digital de textura: Técnicas utilizadas en aplicaciones actuales de CBIR

Abstract

The work that is presented in this article is an overview of the main methods and techniques of visual texture feature for digital image processing, particularly in image retrieval systems based on content (CBIR). This paper also presents three CBIR applications in which the emphasis is on texture features and their techniques used. In CBIR an image is represented by a collection of visual features that describe the content of the image. The primitive features or low-level visual information is color, texture or shape. Texture can be seen as the repetition of a basic spatial pattern, the structure can be periodic or partially periodic. In addition, the texture of a surface refers to the distribution of intensity values in spatial and their properties are evaluated in: fineness, roughness, smoothness, granulation, randomness, etc. There is no standard classification of the methods and techniques for texture analysis, in this review we show that most stands out: a) structural and syntactic methods, and b) Statistical methods. Applications of texture in image retrieval are specific domains such as medical image analysis, remote recognition as aerial images, or industrial inspection.