Credits

LiveComics' panel view is drawn by our own model. It was built with these datasets and open-source tools, and we're grateful to their creators.

Datasets

TINTIN Corpus: Panels

Neil Cohn and the Visual Language Lab, Tilburg University. Panel layout and reading-order tracings, used with permission. doi:10.34894/KSVSRL

Cohn, Neil, Andrew T. Hendrickson, Bruno Cardoso, Bien Klomberg, Irmak Hacımusaoğlu, Ana Krajinović, Sharitha van der Gouw, Fred Atilla, Abe Simons, Tim Hankart, Nanne van Noord, Sam Titarsolej, Fernando Casanova Martínez, Tomás Gaete Altamirano, Marianna Pagkratidou, Michał Szawerna, Nicolas Verstappen, Heinz Insu Fenkl, Leandro Kruszielski, Anna Marta Marini, Gaurav Singh, Dušan Stamenković, Miloš Tasić, and Yen Na Yum. 2026. "The properties of panels in global comics: Frequency and size of 76K panels in 1,030 comics from 144 countries." Language Resources and Evaluation 60 (47):1-24. doi:10.1007/s10579-026-09925-9

Cohn, Neil. 2026. "TINTIN Corpus: Panels, data and preprint." DataverseNL, V2.

Manga109-s

The Aizawa-Yamasaki-Matsui Lab, The University of Tokyo, and the manga artists who allowed their work to be used. Our model was trained using the Manga109-s dataset. manga109.org

Baek, Jeonghun, Atsuyuki Miyai, Shota Onohara, Hikaru Ikuta, and Kiyoharu Aizawa. "Manga109-v2026: Revisiting Manga109 Annotations for Modern Manga Understanding." ICML Workshop, 2026.

Aizawa, Kiyoharu, Azuma Fujimoto, Atsushi Otsubo, Toru Ogawa, Yusuke Matsui, Koki Tsubota, and Hikaru Ikuta. "Building a Manga Dataset 'Manga109' with Annotations for Multimedia Applications." IEEE MultiMedia 27 (2):8-18, 2020. doi:10.1109/mmul.2020.2987895

Matsui, Yusuke, Kota Ito, Yuji Aramaki, Azuma Fujimoto, Toru Ogawa, Toshihiko Yamasaki, and Kiyoharu Aizawa. "Sketch-based Manga Retrieval using Manga109 Dataset." Multimedia Tools and Applications 76 (20):21811-21838, 2017. doi:10.1007/s11042-016-4020-z

COMICS

Iyyer, Mohit, Varun Manjunatha, Anupam Guha, Yogarshi Vyas, Jordan Boyd-Graber, Hal Daumé III, and Larry Davis. "The Amazing Mysteries of the Gutter: Drawing Inferences Between Panels in Comic Book Narratives." IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017.

CoMix: Comic Books Dataset v0.1, Pages (tiny)

Emanuele Vivoli. Public-domain pages from the Digital Comic Museum, CC0 1.0. huggingface.co/datasets/emanuelevivoli/comix-v0_1-pages-tiny

Vivoli, Emanuele, Marco Bertini, and Dimosthenis Karatzas. "CoMix: A Comprehensive Benchmark for Multi-Task Comic Understanding." NeurIPS, 2024.

Comic book Panel Detection

By kristian-wufn3 on Roboflow Universe, version 0.0.2, licensed CC BY 4.0. universe.roboflow.com/kristian-wufn3/comic-book-panel-detection-esufi

Changes made: we used the training split only, kept one un-augmented copy of each page at its original shape, and converted the panel outlines to boxes.

COCO

Our model starts from D-FINE weights pretrained on COCO. COCO annotations are licensed CC BY 4.0 by the COCO Consortium. cocodataset.org

Lin, Tsung-Yi, et al. "Microsoft COCO: Common Objects in Context." European Conference on Computer Vision (ECCV), 2014.

Models and software

D-FINE

The detector architecture and pretrained weights our panel model is built on. Apache License 2.0. github.com/Peterande/D-FINE

Peng, Yansong, Hebei Li, Peixi Wu, Yueyi Zhang, Xiaoyan Sun, and Feng Wu. "D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution Refinement." arXiv:2410.13842, 2024.

Comic text and bubble detector

By ogkalu. Finds speech balloons and captions so no words get cut off at a panel's edge. Apache License 2.0. huggingface.co/ogkalu/comic-text-and-bubble-detector

Runtime

ONNX Runtime, MIT License, Copyright (c) Microsoft Corporation.
NumPy, BSD 3-Clause License, Copyright (c) NumPy Developers.
Pillow, MIT-CMU License, Copyright (c) Jeffrey A. Clark and contributors.
Python, Python Software Foundation License.

Spot a mistake, or made something listed here and want the credit worded differently? Tell us through the inquiries page.