Background
Chinese Modern Poetry is hard to understand, even for Chinese readers. Therefore, we aim to make Chinese Modern Poetry easier to understand by performing basic data analysis and visualization to identify their overall charateristics, and generating one image per poem to assist the reader’s understanding of the poems.
The data source we are using is 小說與詩 provided by editor and writer 魏鵬展, and published by 香港小說學會, 香港詩歌協會 (1st and 2nd issues), and 香港小說與詩協會 (for rest of the issues). The issues of 小說與詩 are indexed in the Hong Kong Literature Database, which is developed by the CUHK Library. We would like to thank CUHK Library for their support in providing this data source.
As the name suggested, 小說與詩 contains both Chinese novels and (modern) Chinese poems. We extracted and used 579 poems across 45 issues that are published between 2013 and 2024 from this data source.
Data Analysis and Visualization
To gather insights into Chinese modern poetry, we analyzed various aspects, including sentiment, word count, linguistic complexity and author details.
The results are presented through visualizations such as word clouds, histograms, correlation matrices, and more.
Image Generation
This part of the project aims to help readers to better grasp the core message of each poem.
We implemented a closed-loop image generation workflow, which includes an evaluator that assesses the quality of the generated images.
If an image is deemed unsuitable or fails to represent the poem accurately, it is regenerated.
Acknowledgement
Members of this project: Ngai Ka Shing (CSCIN/3) and Chong Chun Hin (STAT/3).
The code used in this project can be viewed in this github repository: sam1037/CUHK-DAO
We would like to thank the following people/ organization:
1. 魏鵬展先生 (《小說與詩》主編)
2. Mr. Ryun LEE and Dr. WONG Kwong Cheong for coaching us
3. Qinqin ZHANG, Joe CHAN and other coordinators of this event
4. CUHK Library
The following 2 papers are referenced:
[1] Xu Chen and Di Wu, “Automatic Generation of Multimedia Teaching Materials Based on Generative AI: Taking Tang Poetry as an Example,” IEEE Transactions on Learning Technologies, vol. 17, pp. 1353-1366, 2024, DOI: 10.1109/TLT.2024.3378279.
[2] J. Jiang, Y. Ling, B. Li, P. Li, J. Piao, and Y. Zhang, “Poetry2Image: An Iterative Correction Framework for Images Generated from Chinese Classical Poetry,” arXiv:2407.06196v1 [cs.CV], Jun. 15, 2024.