Humans can easily discriminate not only simple figures but also similar complex textures. At present, a large information processing capability is required for a machine to process such texture image information in a short time on the CRT screen or medical images. It is important to analyze the visual information mechanism for those images. In recent studies of visual search, the texton is considered as a pop‐out discrimination element. There remain a number of problems in the recognition of the texture composed of complex patterns such that the texton in not clear. On the other hand, fractal can be a measure for the complexity, and is related closely to the visual system. This paper presents a recognition experiment and an attention allocation model by simultaneous presentation of multiple images, where a set of textures with known fractal dimensions are used as the image information. The relations are examined quantitatively. As a result, it is shown that a relation can be assumed between the fractal dimension (Dp, DH) of the texture and the allocation of attention fi to the texture. For the set (A) of textures, a multiple regression expression f(A) = 97.6 + 20.1 Dp − 55.3 DH can be applied with a high correlation coefficient. Thus, it is verified that the fractal information can be used as the pop‐out discrimination element in the recognition of the texture.
All Science Journal Classification (ASJC) codes
- Theoretical Computer Science
- Information Systems
- Hardware and Architecture
- Computational Theory and Mathematics