Abstract:

Social media, as a key platform for information dissemination and opinion expression in contemporary society, shapes the unique spatial construction mechanisms and manifestations specific to the information age through its information presentation and user behavior. This paper takes the Librairie Avant-Garde, an avant-garde bookstore located in Nanjing, as a case study. By analyzing images uploaded by users on the Dianping website over the past three years, it explores the characteristics of visual production in cultural spaces. The study employs quantitative methods and utilizes two computer vision techniques—content recognition and object detection—to label the images for in-depth analysis. The research reveals that the visualized cultural space is intricately linked to and distinct from physical space in terms of the authenticity reconstruction of material space, intuitive understanding of cultural symbols, and identity construction of individuals. It demonstrates characteristics of virtual spaces, such as selectivity, symbolic representation, and attitudinal expression. This study applies deep learning technology to the quantitative analysis of image big data, aiming to explore the visual perception of "influencer" spaces and their visual interpretation of complex spatial meanings in the social media era.

Key words: cultural space, bookstore, production of space, social media, computer vision, visual space

摘要：

社交媒体作为当代社会信息流通和观点表达的关键平台，其信息呈现方式和用户行为共同塑造了信息时代特有的空间建构机制及其表现方式。本文选取位于南京的先锋书店（Librairie Avant-Garde）作为案例，通过分析近三年来用户在大众点评网站上传的图片，探讨了文化空间视觉化生产的特征。研究采用量化方法，运用两种计算机视觉技术——内容识别和对象检测——对图片进行标签化处理，并基于此进行深入分析。研究发现，视觉化的文化空间在物质空间原真性的重构、文化符号直观性的理解以及个体的身份构建等方面与现实空间既相互关联又具有独特性，展现出选择性、符号化和态度化等虚拟空间特征。本研究将深度学习技术应用于图像大数据的量化分析，旨在探讨“网红”空间的视觉感知及其在社交媒体时代的空间复合意义与视觉层面的解读。

关键词: 文化空间, 书店, 空间生产, 社交媒体, 计算机视觉, 视觉空间