Presentation Information
[2N6-GS-2x-04]The Proposal for a Personal Book Management System Using Video Analysis with ChatGPT.
〇kanoko Takahashi1, Yuta Takahashi2 (1. TSUMAnotation, 2. Yachiyo Engineering Co., Ltd.)
[[online]]
Keywords:
ChatGPT,Video processing,Text processing,Book management,OCR
Sales of physical books continue to decline. One reason cited is that managing them is more difficult compared to e-books. To make managing books easier and thus encourage more purchases of physical books, a simple system requiring no specialized knowledge is proposed. In the library field, inventory management systems are used to reduce the burden of book inventory checks, but their scale and usage fees make adoption for personal book management impractical. This research focuses on ChatGPT, a type of AI chatbot, due to its no-code/low-code capability for image, text, and video processing. Previous studies attempted OCR on images of book collections using ChatGPT to list book information, but this proved difficult, requiring each image to be photographed with the relevant area isolated. With advancements in AI chatbots enabling video OCR, the feasibility of applying OCR to filmed videos of book collections to generate book information lists was verified. By filming the spines of book collections and instructing ChatGPT to perform OCR, the book data were converted into text and listed in a CSV file.
