融合DE-GWO与SVR的文化意象预测模型

doi:10.11996/JG.j.2095-302X.2023010184

图学学报 ›› 2023, Vol. 44 ›› Issue (1): 184-193.DOI: 10.11996/JG.j.2095-302X.2023010184

融合DE-GWO与SVR的文化意象预测模型

裴卉宁¹(), 邵星辰¹, 谭昭芸¹, 黄雪芹¹, 白仲航¹^,²()

1.河北工业大学建筑与艺术设计学院，天津 300401
2.河北工业大学国家技术创新方法与实施工具工程技术中心，天津 300401

收稿日期:2022-04-15 修回日期:2022-07-04 出版日期:2023-10-31 发布日期:2023-02-16
通讯作者: 白仲航
作者简介:裴卉宁(1986-)，女，讲师，博士。主要研究方向为计算机辅助设计、人因可靠性。E-mail：peihuining@hebut.edu.cn
基金资助:
河北省社会科学基金项目(HB20YS046)

Prediction model of cultural image based on DE-GWO and SVR

PEI Hui-ning¹(), SHAO Xing-chen¹, TAN Zhao-yun¹, HUANG Xue-qin¹, BAI Zhong-hang¹^,²()

1. School of Architecture and Art Design, Hebei University of Technology, Tianjin 300401, China
2. National Technology Innovation Method and Implementation Tool Engineering Technology Center, Hebei University of Technology, Tianjin 300401, China

Received:2022-04-15 Revised:2022-07-04 Online:2023-10-31 Published:2023-02-16
Contact: BAI Zhong-hang
About author:PEI Hui-ning (1986-), lecturer, Ph.D. Her main research interests cover computer-aided design and human reliability. E-mail：peihuining@hebut.edu.cn
Supported by:
Social Science Foundation of Hebei Province(HB20YS046)

摘要/Abstract

摘要：

为更客观准确的量化文化特征与意象间的关系，提出一种融合混合灰狼优化算法(DE-GWO)与支持向量回归(SVR) 的文化意象预测模型。首先，构建以多组意象词汇为基础的响堂山石窟造像的文化特征的意象空间，并利用眼动追踪技术进行文化意象认知实验，获取被试生理认知数据并对其进行单因素方差分析，进而得到文化意象预测模型的眼动指标参数数据集；其次，引入基于DE算法的差分进化策略以弥补GWO搜索过程陷入停滞状态的问题；再次，利用改进后的GWO算法对SVR模型的参数C和g进行寻优；最后利用构建的DE-GWO-SVR模型实现对文化意象认知的预测。为了进一步证明所构建模型的泛化性，采用BP，ABC-SVR和DT等5种模型进行对比实验，结果表明该模型对于文化意象认知的预测效果更好。

关键词: 混合灰狼优化算法, 支持向量回归, 眼动追踪技术, 文化意象预测

Abstract:

In order to more objectively and accurately quantify the relationship between cultural characteristics and imagery, a cultural image prediction model integrating the hybrid gray wolf optimization algorithm (DE-GWO) and support vector regression (SVR) was proposed. First, the image space of the cultural characteristics of the Xiangtangshan Grottoes statues was constructed based on multiple sets of image vocabulary, and cultural image cognition experiments were conducted using eye tracking technology. In doing so, with the subjects' physiological cognition data obtained, the one-way analysis of variance was carried out, and then the eye movement index parameter data set of the cultural image prediction model was obtained. Secondly, the differential evolution strategy based on the DE algorithm was introduced to make up for the problem of GWO search process stagnation. Thirdly, the improved GWO algorithm was used to change the parameter C of the SVR model and g for optimization. Finally, the constructed DE-GWO-SVR model was utilized to realize the prediction of cultural image cognition. In order to further prove the generalization of the constructed model, five models including BP, ABC-SVR, and DT were involved to conduct comparative experiments. The results show that the proposed model could achieve a better prediction effect on cultural image cognition.

Key words: hybrid gray wolf optimization algorithm, support vector regression, eye tracking technology, cultural image prediction

中图分类号:

TP391

裴卉宁, 邵星辰, 谭昭芸, 黄雪芹, 白仲航. 融合DE-GWO与SVR的文化意象预测模型[J]. 图学学报, 2023, 44(1): 184-193.

PEI Hui-ning, SHAO Xing-chen, TAN Zhao-yun, HUANG Xue-qin, BAI Zhong-hang. Prediction model of cultural image based on DE-GWO and SVR[J]. Journal of Graphics, 2023, 44(1): 184-193.

图/表 14

参考文献 27

[1]	CRILLY N, MOULTRIE J, CLARKSON P J. Seeing things: consumer response to the visual domain in product design[J]. Design Studies, 2004, 25(6): 547-577. DOI URL
[2]	SEARCH P. Exploring the dynamic aesthetics of interaction design[M]//Design, User Experience, and Usability. Design Philosophy and Theory. Cham: Springer International Publishing, 2019: 84-97.
[3]	HU Y K, YU S H, QIN S F, et al. How to extract traditional cultural design elements from a set of images of cultural relics based on F-AHP and entropy[J]. Multimedia Tools and Applications, 2021, 80(4): 5833-5856. DOI
[4]	孙华锋, 郑林欣. 基于两轴分析法的标志设计中文化符号构建研究[J]. 包装工程, 2020, 41(22): 258-263, 277.
	SUN H F, ZHENG L X. Construction and research of cultural symbol in logo design based on two-axis analysis method[J]. Packaging Engineering, 2020, 41(22): 258-263, 277 (in Chinese).
[5]	郑林欣, 孙华锋. 高校校徽的文化符号构成与认知研究[J]. 包装工程, 2020, 41(4): 244-251.
	ZHENG L X, SUN H F. Cultural symbol constitution and cognition of badge for universities[J]. Packaging Engineering, 2020, 41(4): 244-251 (in Chinese).
[6]	卢兆麟, 李升波, 徐少兵, 等. 基于眼动跟踪特征的汽车造型评价方法[J]. 清华大学学报: 自然科学版, 2015, 55(7): 775-781.
	LU Z L, LI S B, XU S B, et al. Automobile style evaluation based on eye tracking[J]. Journal of Tsinghua University: Science and Technology, 2015, 55(7): 775-781 (in Chinese).
[7]	汶晨光, 苟秉宸, 吴林健, 等. 基于眼动分析的文化设计基因提取与应用研究[J]. 计算机工程与应用, 2018, 54(11): 217-224, 235. DOI
	WEN C G, GOU B C, WU L J, et al. Extraction and application research of culture design gene based on eye movement analysis[J]. Computer Engineering and Applications, 2018, 54(11): 217-224, 235 (in Chinese). DOI
[8]	闫龙华, 林丽, 肖华亮, 等. 基于心理与生理测量的苗族服饰纹样美感研究[J]. 图学学报, 2017, 38(5): 683-688.
	YAN L H, LIN L, XIAO H L, et al. Research on the aesthetic feeling of Miao costume patterns based on psychological and physiological measurements[J]. Journal of Graphic Science, 2017, 38(5): 683-688 (in Chinese).
[9]	朱月, 邓成连. 基于心理与生理测量的辽瓷文化意象基因研究[J]. 计算机辅助设计与图形学学报, 2020, 32(7): 1183-1191.
	ZHU Y, DENG C L. Study on culture image meme of Liao dynasty ceramics by psychological and physiological measurement[J]. Journal of Computer-Aided Design ＆ Computer Graphics, 2020, 32(7): 1183-1191 (in Chinese).
[10]	SUN L, SHAO H, LI S, et al. Integrated application of eye movement analysis and beauty estimation in the visual landscape quality estimation of urban waterfront park[J]. International Journal of Pattern Recognition and Artificial Intelligence, 2018, 32(9): 1-19.
[11]	李运, 郭钢. 基于多项眼动数据的产品造型方案评选模型[J]. 计算机集成制造系统, 2016, 22(3): 658-665.
	LI Y, GUO G. Selection model of product shape schemes based on multiple eye movement data[J]. Computer Integrated Manufacturing Systems, 2016, 22(3): 658-665 (in Chinese).
[12]	LIU Z, CHEN L, CHANG H, et al. Online algorithms for BP functions maximization[J]. Theoretical Computer Science, 2021, 858: 114-121. DOI URL
[13]	张贇, 吉卫喜, 党英, 等. 基于可拓实例与GA-BP的制动臂成本估算方法[J]. 机械设计与研究, 2021, 37(1): 91-96.
	ZHANG Y, JI W X, DANG Y, et al. Research on product cost estimation method based on extension examples and GA-BP[J]. Machine Design & Research, 2021, 37(1): 91-96 (in Chinese).
[14]	ZHANG H, LIU F, ZHOU Y, et al. A hybrid method integrating an elite genetic algorithm with tabu search for the quadratic assignment problem[J]. Information Sciences, 2020, 539: 347-374. DOI URL
[15]	CURA T. A rapidly converging artificial bee colony algorithm for portfolio optimization[J]. Knowledge-Based Systems, 2021, 233: 1-16.
[16]	TANG H, SUN W, LIN A, et al. A GWO-based multi-robot cooperation method for target searching in unknown environments[J/OL]// Expert Systems with Applications. (2021-08-21) [2022-03-30]. https://doi.org/10.1016/j.eswa.2021.115795.
[17]	王新亭, 邓玮丹. 基于眼部追踪技术的产品意象草图认知差异研究[J]. 计算机辅助设计与图形学学报, 2019, 31(2): 287-294.
	WANG X T, DENG W D. Cognitive differences of product image sketches based on sketch eye tracking[J]. Journal of Computer-Aided Design & Computer Graphics, 2019, 31(2): 287-294 (in Chinese).
[18]	陈田, 樊明焱, 任福继, 等. 采用瞳孔位置实现情感识别的方案[J]. 计算机应用研究, 2021, 38(6): 1765-1769.
	CHEN T, FAN M Y, REN F J, et al. Emotion recognition using pupil position[J]. Application Research of Computers, 2021, 38(6): 1765-1769 (in Chinese).
[19]	刘淼, 聂桂平, 马翀, 等. 感性量化评价技术在工业产品设计中的认知研究与应用: 以智能电源产品设计的眼动实验为例[J]. 东华大学学报: 自然科学版, 2015, 41(4): 536-541.
	LIU M, NIE G P, MA C, et al. Cognition research and application of kansei quantitative evaluation in industrial product design: case study from eye tracking test on design of intelligent power device[J]. Journal of Donghua University: Natural Science, 2015, 41(4): 536-541 (in Chinese).
[20]	ZENG Z, ZHANG M, CHEN T, et al. A new selection operator for differential evolution algorithm[J]. Knowledge-Based Systems, 2021, 226: 1-13.
[21]	SUN G J, XU G N, JIANG N. A simple differential evolution with time-varying strategy for continuous optimization[J]. Soft Computing, 2020, 24(4): 2727-2747. DOI
[22]	唐仲明. 论“响堂样式”的特征及形成[J]. 敦煌研究, 2015(5): 58-63.
	TANG Z M. On the emergence of the style of the Xiangtangshan grottoes and the features of the caves[J]. Dunhuang Research, 2015(5): 58-63 (in Chinese).
[23]	罗仕鉴, 潘云鹤. 产品设计中的感性意象理论、技术与应用研究进展[J]. 机械工程学报, 2007, 43(3): 8-13.
	LUO S J, PAN Y H. Review of theory, key technologies and its application of perceptual image in product design[J]. Chinese Journal of Mechanical Engineering, 2007, 43(3): 8-13 (in Chinese).
[24]	袁树植, 高虹霓, 王崴, 等. 基于感性工学的人机界面多意象评价[J]. 工程设计学报, 2017, 24(5): 523-529. DOI
	YUAN S Z, GAO H N, WANG W, et al. Multi-image evaluation for human-machine interface based on Kansei engineering[J]. Chinese Journal of Engineering Design, 2017, 24(5): 523-529 (in Chinese).
[25]	周爱民, 刘宏斌, 张书涛, 等. 面向汽车主客观审美评价的不确定性推理模型[J]. 浙江大学学报: 工学版, 2021, 55(3): 419-429, 440.
	ZHOU A M, LIU H B, ZHANG S T, et al. Uncertainty reasoning model of subjective and objective aesthetic evaluation of car[J]. Journal of Zhejiang University: Engineering Science, 2021, 55(3): 419-429, 440 (in Chinese).
[26]	程永胜, 徐骁琪, 陈国强, 等. 基于神经网络的电动汽车造型意象预测模型[J]. 计算机集成制造系统, 2021, 27(4): 1135-1145. DOI
	CHENG Y S, XU X Q, CHEN G Q, et al. Image prediction model of electric vehicle based on neural network[J]. Computer Integrated Manufacturing Systems, 2021, 27(4): 1135-1145 (in Chinese).
[27]	GRILLI E, DININNO D, MARSICANO L, et al. Supervised segmentation of 3D cultural heritage[C]//2018 3rd Digital Heritage International Congress Held Jointly with 2018 24th International Conference on Virtual Systems & Multimedia. New York: IEEE Press, 2018: 1-8.

样本编号	刻板的- 生动的	敦厚的- 狡猾的	拘谨的- 豪迈的	粗糙的- 精巧的
1	0.63	-0.74	-0.37	0.42
2	1.37	1.00	1.16	1.00
3	1.11	-0.16	0	0.63
4	0.89	0.32	0.84	0.11
5	0.95	0.21	0.37	1.00
…	…	…	…	…
61	0.84	0.26	-0.42	0.47
62	0.37	-0.21	-0.26	0.89
63	-0.16	-0.37	-0.37	0.83

样本编号	刻板的- 生动的	敦厚的- 狡猾的	拘谨的- 豪迈的	粗糙的- 精巧的
1	0.63	-0.74	-0.37	0.42
2	1.37	1.00	1.16	1.00
3	1.11	-0.16	0	0.63
4	0.89	0.32	0.84	0.11
5	0.95	0.21	0.37	1.00
…	…	…	…	…
61	0.84	0.26	-0.42	0.47
62	0.37	-0.21	-0.26	0.89
63	-0.16	-0.37	-0.37	0.83

眼动指标	F	显著性
总访问持续时间(s)	3.164	0.025
第二次注视时间(s)	251.915	0.000
最大瞳孔直径(mm)	47.462	0.000
平均水平距离(dx)	20.312	0.000
眨眼次数(N)	15.987	0.000
平均眨眼次数(N/s)	497.853	0.000

眼动指标	F	显著性
总访问持续时间(s)	3.164	0.025
第二次注视时间(s)	251.915	0.000
最大瞳孔直径(mm)	47.462	0.000
平均水平距离(dx)	20.312	0.000
眨眼次数(N)	15.987	0.000
平均眨眼次数(N/s)	497.853	0.000

组号					数据集
组号	1	2	3	4	…	24	25	26	27
组1	0.00	-1.00	-1.00	-0.82	…	0.04	0.05	0.16	0.17
组2	-0.49	1.00	1.00	-0.36	…	0.62	-0.13	0.71	0.75
组3	-0.47	-1.00	-1.00	-0.39	…	-0.28	-0.49	-0.10	0.17
组4	-1.00	-1.00	-1.00	0.86	…	-0.22	-0.38	-0.16	0.25
组5	-0.82	-1.00	-1.00	0.82	…	-0.86	-0.34	-0.85	-0.75
组6	0.21	-1.00	-1.00	-0.25	…	-0.77	-0.77	-0.72	-0.33
组7	-1.00	-1.00	-1.00	-0.36	…	0.97	0.61	0.87	0.00
组8	-1.00	-1.00	-1.00	-0.89	…	0.62	0.30	0.32	0.42
组9	0.37	0.05	0.14	0.10	…	0.24	-0.40	-0.47	-0.58
组10	-0.91	-1.00	-1.00	-0.89	…	-0.77	-0.71	-0.66	-0.50
组11	-0.76	-1.00	-1.00	-0.50	…	-0.48	-0.40	-0.46	-0.50
…	…	…	…	…	…	…	…	…	…

融合DE-GWO与SVR的文化意象预测模型

Prediction model of cultural image based on DE-GWO and SVR

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 14

参考文献 27

相关文章 2

编辑推荐

Metrics

本文评价

序号	眼动指标	含义
1	第二次注视时间(s)	指注视点从第二次进入到离开某个区域之间所用的总时长，在对象判别任务中用来衡量被试对对象包含信息的认知与处理程度，并以此反映语义一致性对对象认知的影响
2	总访问持续时间(%)	指某个区域内注视时间的总和所占比，反映被试的认知困难程度。其访问时间比越高则代表被试对相关区域具有更高的关注程度，另一方面也表示对于该区域的信息处理难度也越大
3	平均水平距离(px)	指2个注视点之间间隔的水平距离，表示获取信息量的多少，反映对信息的认知效率和处理难度。间隔距离越大，表明该注视所获取的信息越多且认知效率越高
4	最大瞳孔直径(px)	指样本刺激过程中被试的瞳孔直径最大值，其大小能够反映被试的兴趣程度与情绪的唤醒程度。瞳孔直径越大则代表被试的心理唤醒程度越高，其兴趣程度越高
5	眨眼次数(N)	指实验过程中眨眼次数的总和，除与被试参与认知任务时产生的负荷有关，还与个体的注意力呈负相关。眨眼次数越多，代表被试注意力较为分散，反之则表示相对集中
6	平均眨眼次数(N/s)	指实验过程中眨眼次数的总和与时间的比值，能够反映被试在实验过程中整体注意水平和认知负荷能力

模型	性能指标
模型	MSE	MAE	训练时间(s)
GWO-SVR	0.28	0.39	70.73
ABC-SVR	0.38	1.98	269.86
PSO-SVR	0.41	1.85	204.35
AdaBoost-BP	0.69	0.73	30.15
DT	0.53	0.63	21.02
DE-GWO-SVR	0.21	0.35	75.38

[1]	周爱民1,2，苏建宁 2，阎树田 1，欧阳晋焱 2，张书涛 2. 基于形态美度的产品多意象预测模型[J]. 图学学报, 2018, 39(4): 654-660.
[2]	闫龙华，林丽，肖华亮，时雷. 基于心理与生理测量的苗族服饰纹样美感研究[J]. 图学学报, 2017, 38(5): 683-688.