基于视觉注意模型的洗净率客观评价方法研究  

Objective Evaluation Method of Cleaning Rate Based on Visual Attention Model

作　　者：徐平华 邵芬娟 刘红[2] 丁雪梅[2] 沈岳[3] 

机构地区：[1]浙江理工大学服装学院,服装数字化技术浙江省工程实验室 [2]东华大学服装与艺术学院 [3]南通大学纺织与服装学院

出　　处：《纺织导报》2018年第6期88-90,共3页China Textile Leader

基　　金：国家自然科学基金面上项目(61702460);浙江省教育厅科研项目(Y201738456);南通市科技计划项目(GY12016032);浙江理工大学一流学科优秀博士专项(2017YBZX15);浙江理工大学科研启动基金(17072067-Y)

摘　　要：针对织物污渍残留的评定,文章从人眼视觉注意模型出发,利用污渍图像的亮度、色彩及方向3方面底层特征,由底至上构建污渍视觉热点图,并提出视觉覆盖率和分散度2个指标,综合得出污渍洗净率。结果表明,所提出的污渍洗净率能够反映污渍洗净程度,与人工评级结果显著相关。该方法能够有效描绘出污渍视觉显著程度,可用于纺织品服装的洗净率的客观评级。Currently, the residue of fabric stains is tested by subjective visual measurement rating. Taking advantage of three low-level features of visual attention model, i.e., the brightness, color and direction of the stains image, the visual heat map of stains were created. And combining with visual coverage rate and dispersion, the cleaning rate of stains was obtained comprehensively. Results show that the cleaning rate of stains obtained by the visual attention model is superior to that of subjective visual measurement rating. This method can efectively describe the visual signifcative degree of stains, which can be used for the rating of the cleaning rate of textiles and garment.

关 键 词：视觉注意 热点图 洗净率 客观评级 

分 类 号：TS107[轻工技术与工程—纺织工程]