CoSMo: A Multimodal Transformer for Page Stream Segmentation in Comic Books

Artemis LLabr\'es; Dimosthenis Karatzas; Emanuele Vivoli; Marc Serra Ortega

REVIEW 3 major objections 7 minor 17 references

CoSMo: A Multimodal Transformer for Page Stream Segmentation in Comic Books

T0 review · 3 major / 7 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A compact Transformer that splits comic books into stories, ads, and covers reaches 98.10 F1-Macro on a new 20,800-page benchmark.

desk verdict Useful new task and dataset for comic-book page stream segmentation, but the missing train/test split makes the headline numbers unverifiable as written. read the letter →

arxiv 2507.10053 v1 pith:WYNOXRRO submitted 2025-07-14 cs.CV

Marc Serra Ortega , Emanuele Vivoli , Artemis Llabrés , Dimosthenis Karatzas This is my paper · ORCID

classification cs.CV

keywords PageStreamSegmentationcomicbookunderstandingmultimodalTransformersequencelabelingdocumentlayoutanalysisvision-languagemodelmetadataenrichmentpanopticquality

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper asks whether a small, task-specific Transformer can automatically divide scanned comic books into their semantic sections—stories, advertisements, text stories, and covers—well enough to replace manual metadata annotation. It answers yes: CoSMo, trained on a newly curated 20,800-page dataset of 430 books, reaches 98.10 F1-Macro and 95.08 Panoptic Quality in its multimodal form, and 97.30 F1-Macro in a cheaper vision-only form. The authors report that both variants beat much larger zero-shot vision-language models and traditional layout-feature baselines. The practical point is that the first stage of comic-book content understanding, knowing where one story ends and the next begins, can be automated at scale.

What carries the argument

The load-bearing mechanism is a dual-token Transformer encoder: each page is represented by a frozen pretrained visual embedding and, in the multimodal variant, a text embedding of structured OCR, and the two tokens are interleaved so the encoder can attend over the whole page sequence. Four Transformer layers with absolute positional encoding model narrative flow, and only one token per page feeds a three-layer classification head. A cost-sensitive cross-entropy loss reweights classes by inverse frequency. The same architecture with only visual tokens forms the lightweight vision-only variant; the design point is that decisions come from cross-page context rather than page-local features alone.

What would settle it

Have several independent annotators re-label a random sample of about 50 books with the same protocol, measure agreement specifically on First-Page boundaries, and retrain and evaluate CoSMo on majority-vote labels. If F1-Macro or Panoptic Quality falls materially, or if annotators disagree on a large share of story starts, the reported state of the art is partly an artifact of one annotation protocol.

Watch

Extended reading notes

Core claim

The central discovery is that comic-book page-stream segmentation can be recast as multiclass sequence labeling over a page stream, with each page assigned one of Cover, Advertisement, Text Story, Story, or First-Page, the last being a derived label for the first page of a narrative block. A compact encoder-only Transformer that attends across the whole book solves this task to near-ceiling accuracy on the curated dataset: the multimodal variant reaches 98.10 F1-Macro, 95.08 Panoptic Quality, and a minimum-drags metric of 0.437, while the vision-only variant is nearly as strong. The authors interpret this as evidence that visual layout and style carry most of the segmentation signal, with text helping specifically on ambiguous pages such as ads that resemble stories or first pages with weak title cues. They also position the work as the first method that ingests an entire volume, in contrast to prior single-page or panel-sequence comic understanding work.

Load-bearing premise

The results rest on the assumption that the manually curated, archive-aligned labels, especially the derived First-Page class, are accurate and consistent across all 20,800 pages; quality checks are reported, but no inter-annotator agreement or annotation-error rate is given, so systematic label noise would propagate into every headline number.

Editorial extensions

If this is right

If the reported accuracy holds, archives can auto-generate story-boundary metadata for backlogs of scanned comic books at roughly one percent error instead of tagging pages by hand.
A task-specific model with a small encoder outperforming a much larger zero-shot vision-language model suggests that general-purpose models are not the cheapest route for structured document segmentation.
The strong vision-only results imply that OCR is not required for most pages, so segmentation can run cheaply where text extraction is costly or impossible.
Reliable page-stream segmentation unlocks downstream analysis, character tracking, story indexing, and metadata enrichment, on story blocks rather than on whole issues.
First-Page detection is the hardest class and the main source of residual error; further gains depend on better long-range context modeling for story-boundary cues.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The dataset is drawn from one public-domain Golden Age archive; whether the roughly 98 percent F1 transfers to modern comics, manga, or European albums is untested and would be a natural domain-shift study.
A minimum-drags value below one suggests a human annotator would need fewer than one drag-and-drop correction per book on average, but how sensitive that metric is to label noise is not examined.
A concrete extension would be fine-tuning on a small sample of modern books to measure transfer; the paper's own failure cases already point to atypical first pages and ad-like mixed content as the error clusters to attack.
Since First-Page is a derived label, publishing the exact derivation protocol is what would make the benchmark reproducible; a different story-boundary definition would change labels and headline numbers.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Quick take: this is the first credible formalization of page stream segmentation for comic books, with a genuinely new 430-book/20,800-page dataset and sensible stream-level metrics. But the evaluation never states a train/validation/test split, so the reported SOTA numbers—F1-Macro 98.10, PQ 95.08—could be in-sample. That's not a minor omission; it is the difference between a benchmark you can trust and a claim you can't check.

What's actually good: the task framing with Cover, Advertisement, Text Story, Story, and derived First-Page labels is right and fills a real gap. The dataset, manually curated from public archives and aligned to comics.org, looks like a useful resource even if label quality is only asserted. The architecture is standard—frozen SigLIP features plus a four-layer transformer—but the paper doesn't oversell it. The ablations across backbones and modalities are informative, and the finding that vision dominates while text helps on ambiguous First-Page cases is a fair, non-obvious empirical point. The GitHub release is the right move.

The soft spots are real. Most importantly, Section 6.4 gives learning rate, loss, and early stopping but no split, no seeds, no per-book grouping, no repeated runs. Tables 1–3 report single numbers. Since CoSMo is a transformer over entire book streams, any book appearing in both training and evaluation allows the model to memorize story transitions—exactly the local structure First-Page labels capture. All baselines are evaluated on the same unspecified set, so every comparison inherits this risk. The stress-test note lands.

Smaller issues: no inter-annotator agreement or error rate for the manual labels; the First-Page definition is described casually; the LLM baselines are zero-shot only, which is fine as a comparison but shouldn't be phrased as consistent outperformance. The detection-feature experiment uses a manga-trained detector, which the authors acknowledge may hurt.

Net: the task and dataset are genuinely useful and the paper deserves a careful referee. But I would not accept the results as stated. The authors need to report the split, ideally release a commit-pinned data split and seeds, and rerun the numbers. With that fixed, this becomes a straightforward acceptance.

For you: worth watching, and I'd cite the dataset if the split gets resolved. Bring to reading group if you want a concrete case of why evaluation protocol matters.

Referee Report

3 major / 7 minor

Summary. The paper introduces CoSMo, a transformer-encoder architecture for page stream segmentation (PSS) in comic books, with vision-only and multimodal variants. The authors formalize the task, curate a 430-book/20,800-page dataset labeled into Cover, Advertisement, Text Story, Story, and the derived First-Page classes, and report experiments in which CoSMo outperforms XGBoost, linear probes, zero-shot CLIP/SigLIP, and zero-shot Qwen2.5-VL-32B across F1-Macro, Panoptic Quality (PQ), and the stream-level MnDD metric. The central claim is that a lightweight, task-specific sequential model beats larger general-purpose vision-language models on this task.

Significance. If the reported results withstand scrutiny, CoSMo would be a strong new baseline for comic PSS, and the released dataset would be a useful public resource for a previously absent domain in the PSS literature. The paper's strengths include a clear task formalization, a modular architecture description, systematic ablations over visual backbones and fusion strategies, and the stated intention to release code and data. The finding that visual features dominate while text helps resolve ambiguous cases is plausible and actionable. The main weakness is that the evaluation protocol is incompletely specified, so the generalization claims are currently unverifiable.

major comments (3)

[Section 6.4; Tables 1-3] The paper never specifies the train/validation/test split: it does not state the number of books per split, whether splits are performed at book level, or whether a book used for training can also appear in evaluation. Because CoSMo encodes the full book stream with positional encoding and attention over the entire sequence, any overlap between training and test books would permit memorization of page-transition structure, including the derived First-Page boundaries. All headline numbers (e.g., F1-Macro 98.10, PQ 95.08, MnDD 0.437 in Table 2) therefore lack provenance as measures of generalization. Please state the exact split, guarantee book-level separation, and report results over multiple seeds with mean and standard deviation.
[Section 5; First-Page label] First-Page is described as a derived label marking the first page in a narrative block, but the derivation protocol is not given and no annotation-quality metrics are reported. The paper mentions 'manual annotation' and 'quality checks' without inter-annotator agreement, label error rate, or adjudication details. Since First-Page is consistently the hardest class and stream-level metrics depend on segment boundaries, systematic ambiguity or noise in this derived label propagates into every comparison in Tables 2 and 3. Please provide the annotation protocol and a quantitative assessment of label reliability.
[Section 6.4 and Tables 1-3] All experimental results are single numbers without variance. Given the class imbalance (Section 5) and the modest differences between some variants, such as multimodal versus vision-only CoSMo (F1-Macro 98.10 vs 97.30; PQ 95.08 vs 94.50), error bars or repeated-run statistics are needed to support the claims of consistent improvement and of the multimodal variant being 'best overall.' Report mean and standard deviation over at least three seeds or a book-level bootstrap.

minor comments (7)

[Abstract and Section 1] The abstract and the introductory summary say CoSMo outperforms 'significantly larger general-purpose vision-language models' without consistently adding that the LLM comparisons are zero-shot. Section 6.1 is explicit about zero-shot evaluation, so the high-level statements should carry the same qualifier to avoid overstatement.
[Table 2; Section 6.2] The two Multimodal rows in Table 2 are not labeled by fusion strategy, even though Section 6.2 describes two distinct strategies, fused and multitoken. Add explicit row labels so the reader can map the table to the described variants.
[Equation (4)] The MnDD formula as printed, MnDD = N - sum_{i,j} max_i |Gi ∩ Pj|, is not well-formed because the maximization index appears over the same index used in the summation. Please align the notation with reference [8].
[Section 6.4] The training protocol lists only the learning rate and early stopping. Please also report the optimizer, batch size, number of epochs, and the exact inverse-frequency weighting used in the cost-sensitive cross-entropy loss.
[Section 4.1; Section 7.1] The manuscript uses 'lightweight' to describe CoSMo without giving parameter counts. Add parameter counts for the vision-only and multimodal variants to support the comparison with Qwen2.5-VL-32B.
[Section 7.1] The single-page CoSMo evaluation is reported without describing how a sequence-trained model is applied to an isolated page. Specify whether the page is fed as a one-page sequence, whether context is masked, or whether a separate inference procedure is used.
[Section 1 and Table 2] The introduction states that CoSMo has 'an impressive error rate of approximately 1%,' but Table 2 reports accuracy of 98.65% for the best model, corresponding to a 1.35% error rate. Please reconcile the wording with the reported numbers.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: CoSMo's results come from an externally benchmarked supervised model; self-citations are confined to framing and do not carry the derivation.

full rationale

The paper's central claim is that CoSMo outperforms baselines on comic-book Page Stream Segmentation. The derivation chain is: collect a manually annotated dataset, define a four-class (plus derived First-Page) labeling task, extract frozen SigLIP visual features and Qwen OCR embeddings, train an encoder-only Transformer with cost-sensitive cross-entropy, and evaluate against external baselines (XGBoost, CLIP/SigLIP zero-shot and linear probes, Qwen2.5-VL zero-shot) using metrics adopted from OpenPSS and prior PSS work. No equation in the paper maps fitted parameters to the reported F1-Macro, PQ, or MnDD values; the loss is defined on the same class labels used for evaluation, which is standard supervised learning rather than circularity. The 'First-Page' derived label encodes story boundaries, but that is the task definition, not a smuggled input. The LoCU survey [15] is authored by members of the present team, but it is used only as conceptual framing in the introduction and conclusion and does not determine the architecture, training procedure, or benchmark numbers. The reported results depend on external architectures and a freshly curated dataset, and no self-citation is load-bearing. The missing train/validation/test split and the choice of SigLIP based on the reported test table are reproducibility and selection concerns, not circular reasoning. Therefore, the paper shows no significant circularity.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central claim is an empirical benchmark, so the ledger contains no closed-form derivation constants. The free parameters are disclosed training choices. The load-bearing domain assumptions are the quality of the self-annotated dataset and the transferability of frozen pretrained encoders to comics. No new physical or conceptual entities are introduced; the only new construct is the task formulation itself, which is a definition rather than an invented entity.

free parameters (3)

learning rate = 1e-6
Reported in Section 6.4 as a fixed training choice; no search or sensitivity analysis is reported.
dropout rate = 0.4
Applied throughout the network in Section 4.1.1 as a fixed regularization choice.
cost-sensitive class weights = inverse class frequency
Introduced in Section 6.4 to mitigate the 71% Story-class imbalance; no comparison to other imbalance treatments is provided.

assumptions (5)

domain assumption Frozen SigLIP and Qwen-derived embeddings transfer to comic pages without task-specific adaptation.
Section 4.1.1 relies on pretrained encoders as fixed feature extractors; no evidence is given for transfer beyond the reported results.
domain assumption The manually curated comics.org-aligned labels are accurate, consistent ground truth for story, cover, advertisement, and text-story boundaries.
Section 5 describes thorough quality checks but reports no inter-annotator agreement or error-rate measurement.
domain assumption The derived First-Page label, defined as the first page of a narrative block, is a well-defined and learnable target.
Section 5 introduces First-Page; its consistency across 430 heterogeneous books is asserted, not measured.
domain assumption Cost-sensitive cross-entropy with inverse-frequency weights is an adequate treatment of the class imbalance.
Section 6.4 introduces the weighting scheme; no comparison to alternative imbalance handling is provided.
standard math Standard segment-matching assumptions, such as IoU threshold 0.5, apply to comic page streams.
Section 6.3.2 imports document-level F1, PQ, and MnDD definitions from prior PSS literature without modifications for comic structure.

how reviews work

0 comments

Cite this review

Pith. "Pith review of CoSMo: A Multimodal Transformer for Page Stream Segmentation in Comic Books." pith.science (2026). https://pith.science/paper/WYNOXRRO

@misc{pith2026250710053,
  author       = {Pith},
  title        = {Pith review of: CoSMo: A Multimodal Transformer for Page Stream Segmentation in Comic Books},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WYNOXRRO}},
  note         = {Machine review of arXiv:2507.10053}
}

read the original abstract

This paper introduces CoSMo, a novel multimodal Transformer for Page Stream Segmentation (PSS) in comic books, a critical task for automated content understanding, as it is a necessary first stage for many downstream tasks like character analysis, story indexing, or metadata enrichment. We formalize PSS for this unique medium and curate a new 20,800-page annotated dataset. CoSMo, developed in vision-only and multimodal variants, consistently outperforms traditional baselines and significantly larger general-purpose vision-language models across F1-Macro, Panoptic Quality, and stream-level metrics. Our findings highlight the dominance of visual features for comic PSS macro-structure, yet demonstrate multimodal benefits in resolving challenging ambiguities. CoSMo establishes a new state-of-the-art, paving the way for scalable comic book analysis.

Figures

Figures reproduced from arXiv: 2507.10053 by the authors.

**Figure 2.** Multimodal CoSMo model architecture. are normalized to maintain consistent feature scaling across modalities. For interleaving, we construct a dual-token representation per page—one for the visual modality and one for the textual modality—allowing the Transformer encoder to jointly attend over both representations in sequence. This structure enables the model to capture complementary signals and cross-modal depend… view at source ↗

**Figure 3.** Examples of annotated comic book page types used for [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Example of a comic page sequence showcasing robust and accurate page stream segmentation by both CoSMo Vision and CoSMo [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

**Figure 5.** Figure 5: Examples of pages correctly classified by CoSMo Multi [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

17 extracted references · 16 canonical work pages

[1]

Qwen2.5-vl technical report, 2025

Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, Humen Zhong, Yuanzhi Zhu, Mingkun Yang, Zhao- hai Li, Jianqiang Wan, Pengfei Wang, Wei Ding, Zheren Fu, Yiheng Xu, Jiabo Ye, Xi Zhang, Tianbao Xie, Zesen Cheng, Hang Zhang, Zhibo Yang, Haiyang Xu, and Junyang Lin. Qwen2.5-vl technical report, 2025. 4

2025
[2]

Semantic parsing of interpage relations

Mehmet Arif Demirtas ¸, Berke Oral, Mehmet Yasin Akpınar, and Onur Deniz. Semantic parsing of interpage relations. In 2022 26th International Conference on Pattern Recognition (ICPR), pages 1579–1585. IEEE, 2022. 2

work page 2022
[3]

Document classification and page stream segmentation for digital mailroom applications

Albert Gordo, Marcal Rusinol, Dimosthenis Karatzas, and Andrew Bagdanov. Document classification and page stream segmentation for digital mailroom applications. In Proceed- ings of the International Conference on Document Analysis and Recognition, ICDAR, pages 621–625, 2013. 2

work page 2013
[4]

Large language models for page stream segmentation, 2024

Hunter Heidenreich, Ratish Dalvi, Rohith Mukku, Nikhil Verma, and Neven Piˇculjan. Large language models for page stream segmentation, 2024. 3

work page 2024
[5]

Openpss: An open page stream segmentation benchmark

Ruben van Heusden, Jaap Kamps, and Maarten Marx. Openpss: An open page stream segmentation benchmark. In Linking Theory and Practice of Digital Libraries: 28th In- ternational Conference on Theory and Practice of Digital Li- braries, TPDL 2024, Ljubljana, Slovenia, September 24–27, 2024, Proceedings, Part I , page 413–429, Berlin, Heidel- berg, 2024. Spri...

work page 2024
[6]

Boyd-Graber, Hal Daum ´e, and Larry S

Mohit Iyyer, Varun Manjunatha, Anupam Guha, Yogarshi Vyas, Jordan L. Boyd-Graber, Hal Daum ´e, and Larry S. Davis. The amazing mysteries of the gutter: Drawing in- ferences between panels in comic book narratives. 2017 IEEE Conference on Computer Vision and Pattern Recog- nition (CVPR), pages 6478–6487, 2016. 2

work page 2017
[7]

Panoptic segmentation

Alexander Kirillov, Kaiming He, Ross Girshick, Carsten Rother, and Piotr Doll ´ar. Panoptic segmentation. In Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9404–9413, 2019. 3, 5

work page 2019
[8]

Tab this folder of documents: Page stream seg- mentation of business documents

Thisanaporn Mungmeeprued, Yuxin Ma, Nisarg Mehta, and Aldo Lipani. Tab this folder of documents: Page stream seg- mentation of business documents. 2022. 2, 3, 5

work page 2022

Show all 17 references

[9]

The manga whis- perer: Automatically generating transcriptions for comics

Ragav Sachdeva and Andrew Zisserman. The manga whis- perer: Automatically generating transcriptions for comics. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 12967–12976, 2024. 2

2024
[10]

From panels to prose: Generating literary narratives from comics, 2025

Ragav Sachdeva and Andrew Zisserman. From panels to prose: Generating literary narratives from comics, 2025

2025
[11]

Tails tell tales: Chapter-wide manga transcriptions with char- acter names

Ragav Sachdeva, Gyungin Shin, and Andrew Zisserman. Tails tell tales: Chapter-wide manga transcriptions with char- acter names. In Proceedings of the Asian Conference on Computer Vision (ACCV), pages 2053–2069, 2024. 2

2024
[12]

Comix: A comprehensive benchmark for multi-task comic understanding

Emanuele Vivoli, Marco Bertini, and Dimosthenis Karatzas. Comix: A comprehensive benchmark for multi-task comic understanding. Advances in Neural Information Processing Systems, 37:140828–140846, 2024. 2

2024
[13]

Comics datasets framework: Mix of comics datasets for de- tection benchmarking

Emanuele Vivoli, Irene Campaioli, Mariateresa Nardoni, Niccol`o Biondi, Marco Bertini, and Dimosthenis Karatzas. Comics datasets framework: Mix of comics datasets for de- tection benchmarking. In International Conference on Doc- ument Analysis and Recognition , pages 154–167. ...
[14]

Multimodal transformer for comics text-cloze

Emanuele Vivoli, Joan Lafuente Baeza, Ernest Valveny Llo- bet, and Dimosthenis Karatzas. Multimodal transformer for comics text-cloze. In International Conference on Document Analysis and Recognition, pages 128–145. Springer, 2024. 2

2024
[15]

One missing piece in vision and language: A survey on comics understanding, 2025

Emanuele Vivoli, Mohamed Ali Souibgui, Andrey Barsky, Artemis Llabr´es, Marco Bertini, and Dimosthenis Karatzas. One missing piece in vision and language: A survey on comics understanding, 2025. 2

2025
[16]

Page stream seg- mentation with convolutional neural nets combining textual and visual features

Gregor Wiedemann and Gerhard Heyer. Page stream seg- mentation with convolutional neural nets combining textual and visual features. In Proceedings of the Eleventh Interna- tional Conference on Language Resources and Evaluation (LREC 2018), 2018. 2

2018
[17]

Qwen3 em- bedding: Advancing text embedding and reranking through foundation models, 2025

Yanzhao Zhang, Mingxin Li, Dingkun Long, Xin Zhang, Huan Lin, Baosong Yang, Pengjun Xie, An Yang, Dayiheng Liu, Junyang Lin, Fei Huang, and Jingren Zhou. Qwen3 em- bedding: Advancing text embedding and reranking through foundation models, 2025. 3

2025

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

Qwen2.5-vl technical report, 2025

Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, Humen Zhong, Yuanzhi Zhu, Mingkun Yang, Zhao- hai Li, Jianqiang Wan, Pengfei Wang, Wei Ding, Zheren Fu, Yiheng Xu, Jiabo Ye, Xi Zhang, Tianbao Xie, Zesen Cheng, Hang Zhang, Zhibo Yang, Haiyang Xu, and Junyang Lin. Qwen2.5-vl technical report, 2025. 4

2025

[2] [2]

Semantic parsing of interpage relations

Mehmet Arif Demirtas ¸, Berke Oral, Mehmet Yasin Akpınar, and Onur Deniz. Semantic parsing of interpage relations. In 2022 26th International Conference on Pattern Recognition (ICPR), pages 1579–1585. IEEE, 2022. 2

work page 2022

[3] [3]

Document classification and page stream segmentation for digital mailroom applications

Albert Gordo, Marcal Rusinol, Dimosthenis Karatzas, and Andrew Bagdanov. Document classification and page stream segmentation for digital mailroom applications. In Proceed- ings of the International Conference on Document Analysis and Recognition, ICDAR, pages 621–625, 2013. 2

work page 2013

[4] [4]

Large language models for page stream segmentation, 2024

Hunter Heidenreich, Ratish Dalvi, Rohith Mukku, Nikhil Verma, and Neven Piˇculjan. Large language models for page stream segmentation, 2024. 3

work page 2024

[5] [5]

Openpss: An open page stream segmentation benchmark

Ruben van Heusden, Jaap Kamps, and Maarten Marx. Openpss: An open page stream segmentation benchmark. In Linking Theory and Practice of Digital Libraries: 28th In- ternational Conference on Theory and Practice of Digital Li- braries, TPDL 2024, Ljubljana, Slovenia, September 24–27, 2024, Proceedings, Part I , page 413–429, Berlin, Heidel- berg, 2024. Spri...

work page 2024

[6] [6]

Boyd-Graber, Hal Daum ´e, and Larry S

Mohit Iyyer, Varun Manjunatha, Anupam Guha, Yogarshi Vyas, Jordan L. Boyd-Graber, Hal Daum ´e, and Larry S. Davis. The amazing mysteries of the gutter: Drawing in- ferences between panels in comic book narratives. 2017 IEEE Conference on Computer Vision and Pattern Recog- nition (CVPR), pages 6478–6487, 2016. 2

work page 2017

[7] [7]

Panoptic segmentation

Alexander Kirillov, Kaiming He, Ross Girshick, Carsten Rother, and Piotr Doll ´ar. Panoptic segmentation. In Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9404–9413, 2019. 3, 5

work page 2019

[8] [8]

Tab this folder of documents: Page stream seg- mentation of business documents

Thisanaporn Mungmeeprued, Yuxin Ma, Nisarg Mehta, and Aldo Lipani. Tab this folder of documents: Page stream seg- mentation of business documents. 2022. 2, 3, 5

work page 2022