A Study on Analyzing Symbolism in Literature by Combining Computer Vision Techniques

Abstract

The development of computer technology has greatly promoted the analysis and research of literary works, mainly in the aspects of information collection and data analysis. The analysis model of symbolism in literary works is constructed using the Word2vec algorithm and TextCNN text classification algorithm in this paper. Using computer vision-based OCR text recognition technology, text data is collected from poetic literary works related to a watershed. The dataset is analyzed using the constructed text analysis method to uncover the symbols and symbolic meanings in the sample literary works. Through the analysis of the sample poetic works, two categories of landscape symbols, natural and humanistic, are extracted, and the former is characterized by symbols such as “stream beach”, “shore tide”, “flowers and birds”, “fish and dragons”, etc. Fish and dragons” and other symbols, while the latter is represented by “city buildings” (0.11%), “boats” (0.41%), and “travelers” (0.59%). 0.59%). Five symbolic images are summarized, namely mountain climbing, hearing sounds in temples and monasteries, searching for antiquity in deep forests, boating in rivers and lakes, and fishing songs in idyllic gardens. The research in this paper explores a text-mining method for symbolism in literary works, which is conducive to the development of data-oriented and efficient analysis of literary works.

Keywords

• word2vec
• textcnn
• text mining
• computer vision technology
• literary works
• 97u10