服饰场景视觉共性认知评价方法

Common cognitive evaluation method for fashion visual scenes

针对当前眼动仪在评价多用户视觉聚焦差异方面的缺陷,提出一种基于图像融合的共性视觉热区评价方法。实验依据服饰场景类型,共计选取了60幅典型图像,利用RED SMI桌面式眼动仪对7名受试人员进行实时眼动追踪,记录随机播放条件下的受试者观测视觉场景形成的视觉热区图。分别对同一场景21幅视觉热区二值图进行交、并处理,构建共性视觉热区融合图,并提出了视觉漂移度这个关键性指标。结果表明:所提出的综合视觉漂移度指标可有效剔除个体自身误差,以可视化的方式展现受试者共性视觉聚焦程度,适时弥补了当前多用户视觉聚焦认知评价方法的缺陷。

As the eye tracker lacks the common cognitive evaluation for visual scenes of clothing and accessories, a comprehensive evaluation method of visual hot map based on image fusion was proposed. 7 Subjects′ eye movements were tracked by desktop eye-tracker, and the hot maps were timely recorded while observing 60 different types of representative fashion images. For each vision scene, binary images of hot maps were intersection and overlay processed, and were formed as a visual luminance map. Besides, the visual drift as a critical indicator was proposed. The results show that the proposed visual cognitive indicator can effectively eliminate individual errors, visually display subjects′ common visual focus area, which can fill the blank of current multi-user visual focus cognitive evaluation.