基于眼脑数据和问卷分析的泳装图像情感因子空间构建

Establishment of emotional factor space for swimsuit image based on eye tracking,electroencephalography data and questionnaire analysis

人工智能服装设计中较为重要的步骤是将服装情感设计语言转换为计算机标注语言,而当前研究领域主要是单纯依靠问卷调查结果,这导致标注结果过于主观。针对这一问题,文章以泳装为例,通过发放调查问卷收集主观感性数据,利用眼动仪、脑电系统收集客观生理信号与主观数据进行对比分析,同时验证主观数据的可靠性;对调查问卷数据进行主成分分析和因子分析,最终建立泳装图像情感因子空间。情感因子空间的构建可降低计算机提取服装情感信息特征的难度,有利于实现对泳装图像的识别、分类、检索等后续工作。

An important step in artificial intelligence clothing design is to transform clothing emotional design language into computer annotation language.However,current research mainly relies on questionnaire survey results,which leads to unreliable subjective annotation results.In order to solve this problem,objective physiological signals was collected by eye tracker and electroencephalography(EEG)systems and subjective perceptual data was collect with questionnaire,they were compared and analyzed and the reliability of subjective data was verified.Principal component analysis and factor analysis were carried out on the questionnaire data to establish the final emotional factor space of the swimsuit image.The construction of emotional factor space can reduce the difficulty of extracting emotional information features of clothing by computer,and is conducive to the recognition,classification,retrieval and other follow-up work of swimsuit image.