Abstract:
The low-level visual features(color, texture and shape) in image implicate lots of emotional semantics information with human perception. This paper proposes a new indexing scheme, called Fuzzy Recognize Degree(FRD) clustering using the texture feature. This method allows to retrieve images for emotion-assisted semantic analysis based on high-level affective concepts. For this purpose, a set of perceptual relevant features is introduced: directionality, contrast and coarseness. Collecting indooor decoration images, experimental result shows the high performance of the proposed technique.
Key words:
Fuzzy Recognize Degree(FRD),
emotional semantics,
texture
摘要: 图像的低层视觉特征(颜色、纹理和形状)中包含大量人类可感知的情感语义信息。利用纹理特征,提出一种新的索引方法——模糊认识度(FRD)聚类法,用来描述与情感相关联的语义图像。FRD聚类法能从高层的情感概念出发进行图像检索。索引使用3个感性的纹理特征:方向性,对比度和粗糙度生成FRD值。实验采用室内装饰图片,结果表明,该方法性能较好。
关键词:
模糊认知度,
情感语义,
纹理
CLC Number:
WANG Li. Emotional Semantics Extraction from Image Texture Based on Fuzzy Recognize Degree[J]. Computer Engineering, 2009, 35(20): 212-215.
王 莉. 基于FRD的图像纹理情感语义提取[J]. 计算机工程, 2009, 35(20): 212-215.