Visual interactive meaning evaluation method of movie posters based on deep learning

doi:10.11996/JG.j.2095-302X.2025010221

Abstract

Abstract:

In recent years, the application of deep learning technology for the intelligent evaluation of image aesthetics has become a trend. However, there remains a need for an increase in the amount of annotated data required for high-level aesthetic description tasks, as well as an improvement in the quality and diversity of dataset annotations. To address this, the research took the interactive meaning of visual grammar as a starting point and introduced deep convolutional neural networks for evaluating the visual interactive meaning of movie posters. Firstly, a word segmentation tool was utilized to extract the core semantics of visual interactive meaning from academic literature on movie poster reviews, and the mapping relationship between visual interactive meaning and the characteristic elements of movie posters was summarized, with the aid of morphological analysis. Secondly, a collection of outstanding movie posters was gathered, and a dataset for evaluating the visual interactive meaning of movie posters was constructed in collaboration with expert reviews. Finally, a convolutional neural network was employed to extract features from movie poster samples, establishing a model for evaluating the visual interactive meaning of movie posters, which was verified through practical creation for its feasibility. This method expanded computer aesthetic evaluation to the field of movie poster design, constructing an objective evaluation model by simulating human vision and aesthetic cognition. This model will provide designers with more precise insights into user aesthetic needs and offer references for forward-looking design.

Key words: movie poster, visual grammar, interactive meaning, aesthetic evaluation, deep learning

CLC Number:

WANG Yan, ZHANG Muyu, LIU Xiuzhen. Visual interactive meaning evaluation method of movie posters based on deep learning[J]. Journal of Graphics, 2025, 46(1): 221-232.

Figures/Tables 25

References 18

[1]	KONG S, SHEN X H, LIN Z, et al. Photo aesthetics ranking network with attributes and content adaptation[C]// The 14th European Conference on Computer Vision. Cham: Springer, 2016: 662-679.
[2]	周坤, 张曦, 肖定坤, 等. 基于深度学习的界面设计美学评价研究[J]. 包装工程, 2020, 41(12): 207-215.
	ZHOU K, ZHANG X, XIAO D K, et al. Aesthetic evaluation of interface design based on deep learning[J]. Packaging Engineering, 2020, 41(12): 207-215 (in Chinese).
[3]	CHANG K Y, LU K H, CHEN C S. Aesthetic critiques generation for photos[C]// 2017 IEEE International Conference on Computer Vision. New York: IEEE Press, 2017: 3534-3543.
[4]	金鑫, 周彬, 邹冬青, 等. 图像美学质量评价技术发展趋势[J]. 科技导报, 2018, 36(9): 36-45. DOI
	JIN X, ZHOU B, ZOU D Q, et al. Image aesthetic quality assessment: a survey[J]. Science & Technology Review, 2018, 36(9): 36-45 (in Chinese).
[5]	HOSSAIN N, AHAMAD M M, AKTAR S, et al. Movie genre classification with deep neural network using poster images[C]// 2021 International Conference on Information and Communication Technology for Sustainable Development. New York: IEEE Press, 2021: 195-199.
[6]	李硕. 基于图像信息挖掘的电影推荐算法研究[D]. 长春: 吉林大学, 2020.
	LI S. Research on movie recommendation algorithm based on image information mining[D]. Changchun: Jilin University, 2020 (in Chinese).
[7]	KRESS G, VAN LEEUWEN T. Reading images: the grammar of visual design[M]. London: Routledge, 2006: 45-214.
[8]	王硕, 尹富林. 视觉语法观照下的多模态语篇分析—以一则平面广告为例[J]. 南京工业大学学报(社会科学版), 2011, 10(4): 96-100.
	WANG S, YIN F L. Multimodal advertising discourse analysis from the perspective of visual grammar[J]. Journal of Nanjing Tech University (Social Science Edition), 2011, 10(4): 96-100 (in Chinese).
[9]	宗杰, 赵卫国, 张振江, 等. 基于神经网络的产品文化符号意象认知评价方法研究[J]. 包装工程, 2021, 42(8): 261-267.
	ZONG J, ZHAO W G, ZHANG Z J, et al. Image cognitive evaluation method of product cultural symbols based on neural network[J]. Packaging Engineering, 2021, 42(8): 261-267 (in Chinese).
[10]	曾栋, 周砖. 基于层次化的产品触感评价方法与应用[J]. 包装工程, 2019, 40(18): 123-128, 134.
	ZENG D, ZHOU Z. Evaluation method and application of product tactile perception based on hierarchical structure[J]. Packaging Engineering, 2019, 40(18): 123-128, 134 (in Chinese).
[11]	魏毅东. 论艺术作品的多层次结构[J]. 东南大学学报(哲学社会科学版), 2008, 10(6): 83-88.
	WEI Y D. An analysis of the multi-level structure of artistic works[J]. Journal of Southeast University (Philosophy and Social Science), 2008, 10(6): 83-88 (in Chinese).
[12]	江艳艳. 视觉艺术的具身性多模态话语的认知路径[J]. 牡丹江大学学报, 2019, 28(11): 25-30.
	JIANG Y Y. The cognitive path of multimodal discourse: the embodiment of art[J]. Journal of Mudanjiang University, 2019, 28(11): 25-30 (in Chinese).
[13]	李彦龙, 蔡谦, 孙久康, 等. 基于BP神经网络的汽车外观设计评价方法[J]. 同济大学学报(自然科学版), 2021, 49(1): 116-123.
	LI Y L, CAI Q, SUN J K, et al. An evaluation method of vehicle exterior design based on BP neural network[J]. Journal of Tongji University (Natural Science Edition), 2021, 49(1): 116-123 (in Chinese).
[14]	姚干勤, 薛澄岐. 产品造型风格意象认知空间构建方法研究[J]. 包装工程, 2018, 39(12): 100-106.
	YAO G Q, XUE C Q. Methods of constructing the cognitive space of product styling image[J]. Packaging Engineering, 2018, 39(12): 100-106 (in Chinese).
[15]	黄佳美, 张伟彬, 熊官送. 基于深度卷积神经网络的汽车图像分类算法与加速研究[J]. 现代电子技术, 2024, 47(7): 140-144.
	HUANG J M, ZHANG W B, XIONG G S. Research on automobile image classification algorithm and acceleration based on deep convolutional neural network[J]. Modern Electronics Technique, 2024, 47(7): 140-144 (in Chinese).
[16]	毕春艳, 刘越. 基于深度学习的视频人体动作识别综述[J]. 图学学报, 2023, 44(4): 625-639. DOI
	BI C Y, LIU Y. A survey of video human action recognition based on deep learning[J]. Journal of Graphics, 2023, 44(4): 625-639 (in Chinese).
[17]	熊强, 郑建华. 基于机器学习的平面公益广告效果评价研究[J]. 包装工程, 2023, 44(20): 451-459.
	XIONG Q, ZHENG J H. Effect evaluation of plane public service advertisement based on machine learning[J]. Packaging Engineering, 2023, 44(20): 451-459 (in Chinese).
[18]	邵俊棋, 钱文华, 徐启豪. 基于条件残差生成对抗网络的风景图生成[J]. 图学学报, 2023, 44(4): 710-717. DOI
	SHAO J Q, QIAN W H, XU Q H. Landscape image generation based on conditional residual generative adversarial network[J]. Journal of Graphics, 2023, 44(4): 710-717 (in Chinese).

结构	形态分类	编号
结构	形态分类	1	2	3	4	5	6
媒介层	色系L1	暖色系	冷色系	-	-	-	-
	饱和度L2	高饱和	低饱和	-	-	-	-
	明度L3	明度高	明度低	-	-	-	-
	图像内容L4	人物正面	人物侧面	人物背面	静物	自然环境	无图像
	景别L5	特写	近景	中景	远景	全景	-
技艺层	表现手法L6	抽象	具象	夸张	变形	透视	对比
技艺层	布局结构L7	满版型	对称型	分割型	倾斜型	曲线型	重心型

结构	形态分类	编号
结构	形态分类	1	2	3	4	5	6
媒介层	色系L1	暖色系	冷色系	-	-	-	-
	饱和度L2	高饱和	低饱和	-	-	-	-
	明度L3	明度高	明度低	-	-	-	-
	图像内容L4	人物正面	人物侧面	人物背面	静物	自然环境	无图像
	景别L5	特写	近景	中景	远景	全景	-
技艺层	表现手法L6	抽象	具象	夸张	变形	透视	对比
技艺层	布局结构L7	满版型	对称型	分割型	倾斜型	曲线型	重心型

维度	分类	关键形态	电影海报特征编码
接触	索取	L4	L4 (1)
接触	提供	L4	L4 (2)，L4 (3)，L4 (4)， L4 (5)，L4 (6)
距离	近距离	L5	L5 (1)，L5 (2)，L5 (3)
距离	远距离	L5	L5 (4)，L5 (5)
视角	仰视视角	L6，L7	L6 (3)，L6 (4)，L7 (4)
	平视视角		L6 (2)，L7 (1)，L7 (6)
	俯视视角		L6 (3)，L6 (4)，L7 (4)
情态	高情态	L1，L2，L3	L1 (1)，L2 (1)，L3 (1)
情态	低情态	L1，L2，L3	L1 (2)，L2 (2)，L3 (2)

维度	分类	关键形态	电影海报特征编码
接触	索取	L4	L4 (1)
接触	提供	L4	L4 (2)，L4 (3)，L4 (4)， L4 (5)，L4 (6)
距离	近距离	L5	L5 (1)，L5 (2)，L5 (3)
距离	远距离	L5	L5 (4)，L5 (5)
视角	仰视视角	L6，L7	L6 (3)，L6 (4)，L7 (4)
	平视视角		L6 (2)，L7 (1)，L7 (6)
	俯视视角		L6 (3)，L6 (4)，L7 (4)
情态	高情态	L1，L2，L3	L1 (1)，L2 (1)，L3 (1)
情态	低情态	L1，L2，L3	L1 (2)，L2 (2)，L3 (2)

视觉互动意义四维度	分类	初级评价语义
接触	索取	有目光的、共情的
接触	提供	旁观者视角的、叙事的
距离	近距离	亲近的、特写的
距离	远距离	疏远的、空间感的
视角	仰视视角	强势的、高大的
	平视视角	平等的、亲切的
	俯视视角	弱势的、渺小的
情态	高情态	激情的、热烈的
情态	低情态	灰暗的、冷淡的