Program / GS-07: OCR & Japanese Document/Visual Reading

GS-07: OCR & Japanese Document/Visual Reading

November 05 (Thursday) · 14:00–15:30 · Room 2

Chair: Takeru Aoki (The University of Tokyo)

Back to timetable

14:00–14:20Paper ID: 51Full paper

Graph-Based Character Candidate Selection for Rotated Japanese Character String Recognition

Fumito Nakagawa (Mie University)*; Tetsushi Wakabayashi (Mie University); Kento Morita (Mie University)

14:20–14:40Paper ID: 106Full paper

Separation-Based Recognition of Overlapped Printed Digits Using a Two-Branch Decoder U-Net

Kazuyuki Nakazawa (Mie University)*; Tetsushi Wakabayashi (Mie University); Kento Morita (Mie University)

14:40–15:00Paper ID: 130Full paper

Data Augmentation for Low-Frequency Kuzushiji Character Recognition Using Diffusion Models and Flow Matching

Hiraku Yamada (Mie University)*; Kento Morita (Mie University); Tetsushi Wakabayashi (Mie University)

15:00–15:20Paper ID: 259Full paper

Evaluating Generated Raw Gaze Through Panel Reading Order in Manga

Yuma Iwamoto (University of Electro-Communications)*; Shogo Matsuno (University of Electro-Communications)

Presentations finish at 15:20. The scheduled session ends at 15:30.

Full paper: 15 minutes presentation + 5 minutes Q&A. Short, abstract-only and late breaking papers: 10 minutes presentation + 5 minutes Q&A. Speakers should bring their own laptop; presentation PCs are not provided.