Author(s)

Xiaomeng Zhanag, Haibo Zhang

Corresponding Author

Haibo Zhang

ABSTRACT

In order to recognize and retrieve the emotional semantics of male T-shirt images, it is necessary to extract the low-level features of the color and texture of Images and analyze the emotional semantics to find out the relationship between the image features and emotional semantics. To study the male T-shirt image emotional description and establish the three-dimensional emotional factor space, on the basis of this, the corresponding relationships among color, texture features and three factors of T-shirt images are analyzed. The results show that the first factor can be explained by 10-dimensional features of saturation-cold and warm fuzzy histogram plus 1-dimentional contrast, the second factor can be represented a feature of 256-dimentional gray scales plus 1-dimentional color contrast, the third factor can be used 3-dimentional Tamura features (coarseness, contrast, directionality)to characterize. This relationship lays the foundation for achieving emotional semantic recognition and retrieval of content-based T-shirt images.

KEYWORDS

Male T-shirt image, low-level feature extraction, emotional semantic, color feature, texture feature