基于眼动模式的木质民居个性化定制构件配置方法研究

木质民居是苗侗族传统建筑文化，个性化定制乃是当今木质民居建筑设计的潮流。为快速配置木质民居个性化定制构件，使得其造型意象真正满足消费者隐性需求，提出一种基于眼动模式的个性化定制构件配置方法。首先，研究了木质民居的组成和人们对构件配置的认知。其次，依据蚁群优化算法思想，建立木质民居个性化定制构件配置过程图解模型；将木质民居个性化定制构件库系统中的构件区域划分为兴趣区域（area of interest，AOI）和注视区域（area of fixation，AOF）；以眼动模式中热点图数据为基础，建立眼动数据可视化模型，并以围栏为例，提取注视区域（AOF）的灰度直方图、注视点数目、注视时间及平均注视时间等。再次，计算木质民居个性化定制构件系统中兴趣区域内设计特征的选择概率，以此构建定制构件的适应度函数，寻求配置方案的最优解。最后，以木质民居个性化定制项目为例，论述该方法的配置过程；结合七等级Likert量表实验对眼动实验结果进行对比验证。T检验结果验证了该个性化定制构件配置方法的可行性，能在符合用户隐性需求的基础上进行高效配置，为产品个性化定制模块配置提供一种新的技术方法。

Research on configuration method of wooden residence personalized customization component based on eye movement mode

Wooden residences are the traditional architectural culture of Miao and Dong nationality, and the personalized customization is the trend of wooden residence architecture design. In order to quickly configure the personalized customization component of wooden residence and make its model image truly meet the implicit needs of consumers, a personalized customization component configuration method was presented based on eye movement mode. Firstly, the compositions of wooden residence and people's cognition for component configuration were studied. Secondly, according to the idea of the ant colony optimization algorithm, a graphical model of wooden residence personalized customization component configuration process was established. The component areas in the personalized customization component library system of wooden residences were divided into AOI (area of interest) and AOF (area of fixation). Based on the data of the heat maps in the eye movement mode, the eye movement data visualization model was established. The gray-scale histogram of the area of fixation (AOF), the number of fixation points, the fixation time and the average fixation time were extracted from the example of railing part. Then, the probability selection for design feature in interest area of wooden residence personalized customization component system was calculated and the fitness function of the customized component was constructed to quest the optimal solution of the configuration scheme. Finally, taking the wooden residence customization project as an example, the configuration process of this method was related and analyzed. Combining with 7 level Likert scale experiment,the eye movement test results were compared and validated. The result of the T-test confirmed the feasibility of personalized customization component configuration method which could be configured efficiently on the basis of the implicit requirements of users. This method provides a new technical approach to personalized customization module configuration.