Pith. sign in

REVIEW 4 major objections 8 minor 59 references

DocTTT: Test-Time Training for Handwritten Document Recognition Using Meta-Auxiliary Learning

T0 review · 4 major / 8 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read The central claim of DocTTT is that test-time adaptation, powered by a meta-learned masked-autoencoder auxiliary task, makes handwritten document recognition more accurate by letting each model instance adjust to the specific input page…

desk verdict Applies a known meta-auxiliary test-time training recipe to handwritten document recognition with small, statistically unvalidated gains and somewhat inconsistent ablations; plausible but needs rigor before the 'significant' claim can stand. read the letter →

arxiv 2501.12898 v1 pith:HUYQCCGK submitted 2025-01-22 cs.CV

Wenhao Gu , Li Gu , Ziqiang Wang , Ching Yee Suen , Yang Wang This is my paper · ORCID
classification cs.CV
keywords handwrittendocumentrecognitiontest-timetrainingmeta-learningmaskedautoencodermeta-auxiliarylearninglayoutself-supervisedadaptation
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

DocTTT claims that a handwritten document recognition model becomes more accurate if, at test time, it first adapts itself to the specific input image before reading the page. The adaptation is driven by a self-supervised masked-autoencoder reconstruction loss, so no additional labels are needed during testing. To make this adaptation effective, the model is trained with a meta-learning objective that optimizes the parameters to improve the primary recognition task after one auxiliary gradient step. On the READ 2016, IAM, and RIMES 2011 benchmarks, the paper reports lower character error rates than previous state-of-the-art systems, and its ablations indicate that both the test-time adaptation and the meta-training loop contribute to the gains.

What carries the argument

The carrying mechanism is the Meta-Auxiliary training loop, a bi-level optimization built on MAML. The inner loop adapts the full parameter set $\Theta$ to a masked input by $\Theta' \leftarrow \Theta - \beta \nabla_\Theta \mathcal{L}_{\text{aux}}(x_i;\Theta)$, where $\mathcal{L}_{\text{aux}} = 1 - \text{SSIM}(x_i, \text{reconstruction})$; the outer loop then updates $\Theta$ so that the adapted model minimizes the primary cross-entropy loss $\mathcal{L}_{\text{pri}}(x_i, y_i; \Theta')$ on the ground-truth token sequence. This couples a self-supervised MAE reconstruction branch to the transcription decoder through a shared fully convolutional encoder, and the same inner-loop update is replayed on each test image.

What would settle it

On a held-out set of pages written by authors and in styles absent from training, apply the DocTTT test-time update to each page and compare character error rate against the same model without the update; if the adapted predictions are not consistently better, the claimed benefit of per-input adaptation collapses.

Watch

Extended reading notes

Core claim

The paper sets out to establish that test-time training, applied per document image, lets an HDR model handle diverse handwriting styles and layouts better than a fixed model can. The mechanism is a two-branch network: a fully convolutional shared encoder, a transformer decoder that predicts a flattened XML sequence of text and layout tokens, and an auxiliary masked-autoencoder branch that reconstructs a masked input with an SSIM loss. Training uses a MAML-style bi-level loop: an inner step updates the parameters by the auxiliary MAE loss on a masked training image, and an outer step updates the original parameters so that the adapted model minimizes the primary cross-entropy loss on the ground-truth tokens. At test time the same inner step is applied to a new input without its label, and the adapted model produces the transcription. The reported results are 3.18% CER on READ 2016 single pages, 3.47% on double pages, 4.22% on IAM paragraphs, and 2.33% on RIMES 2011 lines, with ablations showing that removing the test-time adaptation or the meta-learning both degrade accuracy.

Load-bearing premise

The load-bearing premise is that a reconstruction update made on one masked test image moves the shared visual features in a direction that also helps the transcription task; the paper assumes this alignment rather than proving it.

Editorial extensions

If this is right

  • Handwritten document recognition can be improved at test time without any additional ground-truth annotation, since the auxiliary reconstruction loss is self-supervised.
  • The meta-trained model is prepared for one-step adaptation, so the same parameters transfer across page-level, double-page, paragraph, and line-level inputs with fixed hyperparameters.
  • The improvement is not confined to text; layout metrics such as LOER and mAPCER also improve on READ 2016, indicating that the adapted visual features help structural understanding as well.
  • Each component, including positional encoding, teacher forcing, curriculum dropout, test-time training, and meta-learning, contributes to the final accuracy, and removing any of them raises error rates on at least some benchmarks.

Reading between the lines

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

  • An untested extension of the same loop is to swap the masked-autoencoder auxiliary task for other self-supervised objectives, such as rotation prediction, contrastive learning, or inpainting, that force the shared backbone to encode the test page's visual identity; if the benefit is really about per-input adaptation, those objectives should also transfer.
  • Since handwriting style is shared across pages from one writer, adapting once on a small batch of pages by the same author, instead of per page, could capture the same style signal at lower test-time cost; the paper does not investigate this.
  • The paper motivates the method with scarce annotated data but evaluates on full benchmark splits; training on progressively smaller labeled fractions would directly test whether the test-time adaptation compensates for missing supervision.
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.

Referee Report

4 major / 8 minor

Summary. The paper proposes DocTTT, a test-time training framework for handwritten document recognition (HDR). DocTTT uses a masked autoencoder (MAE) as a self-supervised auxiliary task; at test time, it adapts the shared visual backbone to each input image with a few gradient steps on the MAE reconstruction loss, and the model parameters are trained with a bi-level meta-learning objective so that this adaptation improves the primary HDR task. Experiments on READ 2016, IAM, and RIMES 2011 report lower CER and WER than several prior systems, with additional metrics for layout understanding (LOER, mAPCER). The paper includes ablations removing components such as TTT, meta-learning, positional encoding, and teacher forcing.

Significance. If the reported gains are reproducible and statistically robust, DocTTT would be a practical contribution: it shows that per-input adaptation with a self-supervised proxy can improve full-page HDR without extra annotations. The methodological combination of MAML with a masked-autoencoder auxiliary branch is reasonable and addresses a real limitation of fixed models for diverse handwriting styles. However, the current evidence is incomplete: there are no error bars or significance tests, one table directly contradicts the discussion's claim of superiority, the ablations do not cleanly isolate the proposed mechanism, and key hyperparameters are unreported. The central claim is defensible but not yet convincingly established.

major comments (4)
  1. [§4.3, Tables 1–4] The abstract and conclusion state that DocTTT 'significantly outperforms' existing state-of-the-art approaches, but every reported number is a single-run point estimate without error bars, confidence intervals, or significance tests. The margins over the strongest baselines are small (CER 3.18 vs 3.43 on READ 2016 single page, 4.22 vs 4.45 on IAM, 2.33 vs 2.63 on RIMES 2011). Without uncertainty quantification, the central empirical claim cannot be distinguished from run-to-run variation. Please report results over multiple seeds (at least three) with means and standard deviations, or provide a significance test such as paired bootstrap across test examples.
  2. [§4.3, Table 2 and Discussion] On the READ 2016 double-page level, DocTTT's LOER is 3.35%, which is worse than Faster DAN's 3.08%. Yet the Discussion states that DocTTT shows 'superior performance in both CER and WER (for text accuracy) and LOER and mAPCER'. This is factually incorrect for LOER. Moreover, Table 5 shows that 'DocTTT w/o meta-learning' achieves a better double-page LOER (3.27) than the full DocTTT (3.35), contradicting the claim that the meta-learning component improves layout ordering. Please correct the overclaim or provide an explanation (e.g., a footnote that LOER is not statistically compared).
  3. [§4.4, Table 5] The ablation results do not consistently isolate the contribution of the meta-auxiliary mechanism. On IAM, both 'DocTTT w/o TTT' (4.68) and 'DocTTT w/o meta-learning' (4.71) are worse than the Baseline (4.55), while the full method reaches 4.22; on READ 2016 single page, 'DocTTT w/o TTT' (3.48) is essentially identical to Baseline (3.49). This pattern is inconsistent with the claim that each component has 'a positive impact' in isolation. The full-method gain could be due to interaction effects or to run-to-run variation. Please add more controlled variants (e.g., baseline + MAE-only regularization without meta, baseline + meta without MAE, TTT without meta) and repeated runs to disentangle the effects.
  4. [§3.2, §3.3, and §4.3] The paper does not report the key hyperparameters of the proposed method: the inner-loop learning rate β, the meta-learning rate α, the MAE masking ratio, the number of inner-loop adaptation steps, the meta batch size B (used in Eq. 3), and the curriculum schedule (the 90%→20% synthetic-to-real ratio). The statement in §4.3 that 'the same set of hyperparameters are used across all datasets' cannot be verified without these values. Please provide a complete hyperparameter table or an appendix with the exact training configuration, including the number of outer-loop steps, optimizer settings, and data augmentation details.
minor comments (8)
  1. [Table 2 caption] The word 'Evaluation' is misspelled as 'Evalutation'.
  2. [Table 5 caption] The caption says 'Ablation study of on the test set'; remove 'of'.
  3. [Throughout] The dataset name is written inconsistently as both 'Rimes 2011' and 'RIMES 2011'; please standardize to 'RIMES 2011'.
  4. [References] Several references appear unverifiable or placeholder-like, for example [3] 'Lucas E. Anderson and Sofia I. Martinez' at ICML 2022, [27] 'Michael C. Johnson and Emily D. Richards' in Pattern Recognition Letters, [42] 'M Rajalakshmi, P Saranya, and P Shanmugavadivu', and [55] 'David L. Williams and Neha S. Patel' in IEEE TPAMI. Please verify that all bibliography entries correspond to real, published works and provide complete author lists and venues.
  5. [Eq. (5)] The notation '$y \text{ len}_i$' is hard to read; use $\mathrm{len}(y_i^{\text{text}})$.
  6. [Eq. (2)] The constants $a$ and $b$ in the SSIM formula are not defined; typically they are $(K_1 L)^2$ and $(K_2 L)^2$ with $L=255$ and small $K_1, K_2$. Please specify their values.
  7. [Figure 7] The caption says the figure 'demonstrates the effectiveness' of DocTTT, but it only shows a few qualitative examples. Please soften the caption or add quantitative support (e.g., error reduction on those examples).
  8. [§4.3] The sentence 'The results are illustrated in Figures 6 and 7' is vague; clarify which specific quantitative results each figure supports, since Figure 6 is a qualitative layout example and Figure 7 shows transcription corrections.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: DocTTT's central claim is supported by held-out benchmark evaluations and a standard meta-learning objective; self-citations are motivational, not load-bearing.

full rationale

The paper's derivation chain is not circular. The method defines a supervised primary loss Lpri (Eq. 4) and a self-supervised masked-autoencoder loss Laux (Eq. 2). The meta-training objective in Eq. 3 optimizes the original parameters so that inner-loop adaptation via Laux improves Lpri on training instances, which is a standard MAML-style bi-level formulation. At test time, only the self-supervised Laux is used to adapt the model; no test labels are used. The reported claims are evaluated on external public benchmarks (READ 2016, IAM, RIMES 2011) against published baselines, so the central empirical claim is independently falsifiable rather than being equivalent to any fitted input. The self-citations to prior meta-auxiliary works [12,32,59] motivate the architecture and training strategy, but the paper does not rely on those works as proof of the current method's effectiveness; the ablations in Table 5 and comparisons in Tables 1-4 carry that burden. The absence of variance estimates or significance tests is a legitimate concern about evidential strength, but that is a statistical robustness issue, not circularity. No step in the claimed derivation reduces, by construction, to its own inputs.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The central claim rests on several hand-chosen hyperparameters that are not fully specified, plus a set of domain assumptions about the usefulness of MAE as an auxiliary task and the stability of test-time adaptation. No new physical entities or theoretical constructs are introduced.

free parameters (6)
  • Inner-loop learning rate beta = not specified
    Controls the size of test-time adaptation updates in Eq. (1); central to the method.
  • Meta-learning rate alpha = reported as 10^-4 for Adam, but not explicitly linked to alpha or beta
    Used in outer loop Eq. (3); not clearly separated from general optimizer LR.
  • MAE masking ratio = not specified ('a certain percent')
    Determines how much of the input is hidden for the auxiliary reconstruction task.
  • Number of inner-loop adaptation steps = not specified ('a few gradient updates')
    Number of gradient updates at test time affects adaptation quality and compute.
  • Meta batch size B = not specified
    Batch size in Eq. (3) for meta-objective; affects stability.
  • Curriculum schedule for synthetic-to-real ratio = 90% synthetic to 20% over time, exact schedule not given
    Phase 2 training schedule; impacts final model.
assumptions (6)
  • standard math Gradient descent and backpropagation optimize the meta-objective
    Assumes standard differentiability and optimization.
  • domain assumption The MAML update in Eq. (1) with one or few gradient steps is a valid approximation for adaptation
    Relies on [20] and prior meta-auxiliary work [12].
  • ad hoc to paper SSIM is a suitable reconstruction loss for handwritten images
    The paper argues SSIM preserves perceptual quality but does not provide a quantitative comparison for HDR.
  • ad hoc to paper MAE reconstruction is an effective auxiliary task for the HDR primary task
    Central to TTT; assumed without theoretical justification, supported only by ablation.
  • domain assumption Test-time adaptation on a single unlabeled image is stable
    No catastrophic forgetting or variance analysis is provided.
  • domain assumption IAM dataset split avoids train/test overlap for line-level pretraining
    Section 4.3 states a 'similar split' was used, but the exact split and overlap checks are not detailed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DocTTT: Test-Time Training for Handwritten Document Recognition Using Meta-Auxiliary Learning." pith.science (2026). https://pith.science/paper/HUYQCCGK

@misc{pith2026250112898,
  author       = {Pith},
  title        = {Pith review of: DocTTT: Test-Time Training for Handwritten Document Recognition Using Meta-Auxiliary Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HUYQCCGK}},
  note         = {Machine review of arXiv:2501.12898}
}
read the original abstract

Despite recent significant advancements in Handwritten Document Recognition (HDR), the efficient and accurate recognition of text against complex backgrounds, diverse handwriting styles, and varying document layouts remains a practical challenge. Moreover, this issue is seldom addressed in academic research, particularly in scenarios with minimal annotated data available. In this paper, we introduce the DocTTT framework to address these challenges. The key innovation of our approach is that it uses test-time training to adapt the model to each specific input during testing. We propose a novel Meta-Auxiliary learning approach that combines Meta-learning and self-supervised Masked Autoencoder~(MAE). During testing, we adapt the visual representation parameters using a self-supervised MAE loss. During training, we learn the model parameters using a meta-learning framework, so that the model parameters are learned to adapt to a new input effectively. Experimental results show that our proposed method significantly outperforms existing state-of-the-art approaches on benchmark datasets.

Figures

Figures reproduced from arXiv: 2501.12898 by the authors.

Figure 1
Figure 1. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Illustration of the capability of Masked Autoencoders. MAE effectively tackles the HDR problem by reconstructing diverse handwriting styles from unseen test data. The image features four distinct writing styles (first column), a 75% masked image (second column), and reconstructions from both low-resolution (Reconstructed A) and high-resolution (Reconstructed B) inputs. Accurately pre￾dicting token sequences from the… view at source ↗
Figure 3
Figure 3. Illustration of our model architecture. On the left, the input document image is masked and then passes through shared weights to the auxiliary branch for the reconstruction task. We use the auxiliary branch to fine-tune the model for each instance using a self￾supervised masked autoencoder loss. The adapted model is then used for the primary task of predicting the flattened XML representation (see [PITH_FULL_IMAGE… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Illustration of the Meta-Auxiliary training. For a training example xi with the ground-truth label yi, we update the model parameters Θ using a self-supervised MAE loss (inner loop) to obtain an adapted model Θ ′ for this training example. We then use the adapted model…
Figure 5
Figure 5. Figure 5: Illustration of the auxiliary branch. Given an input im [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 7
Figure 7. Figure 7: Comparison of handwriting transcription correc￾tions by DocTTT across four distinct writing styles. This fig￾ure shows original handwriting samples (left), the initial prediction marked with errors highlighted in red (center), and the corrected prediction after applyin…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 39 canonical work pages

  1. [3]

    Anderson and Sofia I

    Lucas E. Anderson and Sofia I. Martinez. Applying trans- formers to handwritten text recognition: A new approach. In Proceedings of the International Conference on Machine Learning, pages 785–793, 2022. 2

    2022

  2. [18]

    Improving cnn-rnn hybrid networks for handwrit- ing recognition

    Kartik Dutta, Praveen Krishnan, Minesh Mathew, and CV Jawahar. Improving cnn-rnn hybrid networks for handwrit- ing recognition. In 2018 16th International Conference on Frontiers in Handwriting Recognition (ICFHR) , pages 80–

    work page 2018

  3. [55]

    Williams and Neha S

    David L. Williams and Neha S. Patel. A comprehensive re- view of large-scale datasets for offline handwritten document recognition. IEEE Transactions on Pattern Analysis and Ma- chine Intelligence, 45(2):567–579, 2023. 2

    2023

  4. [1]

    Improved handwritten digit recognition using convolutional neural networks (cnn)

    Savita Ahlawat, Amit Choudhary, Anand Nayyar, Saurabh Singh, and Byungun Yoon. Improved handwritten digit recognition using convolutional neural networks (cnn). Sen- sors, 20(12):3344, 2020. 2

    2020

  5. [2]

    An improved faster-rcnn model for handwritten character recognition

    Saleh Albahli, Marriam Nawaz, Ali Javed, and Aun Irtaza. An improved faster-rcnn model for handwritten character recognition. Arabian Journal for Science and Engineering , 46(9):8509–8523, 2021. 1

    work page 2021

  6. [4]

    Mt3: Meta test-time training for self- supervised test-time adaption

    Alexander Bartler, Andre B ¨uhler, Felix Wiewel, Mario D¨obler, and Bin Yang. Mt3: Meta test-time training for self- supervised test-time adaption. In International Conference on Artificial Intelligence and Statistics , pages 3080–3090. PMLR, 2022. 2

    work page 2022

  7. [5]

    Docsynth: a layout guided approach for controllable docu- ment image synthesis

    Sanket Biswas, Pau Riba, Josep Llad ´os, and Umapada Pal. Docsynth: a layout guided approach for controllable docu- ment image synthesis. In International Conference on Doc- ument Analysis and Recognition , pages 555–568. Springer,

  8. [6]

    Boosting modern and historical handwrit- ten text recognition with deformable convolutions

    Silvia Cascianelli, Marcella Cornia, Lorenzo Baraldi, and Rita Cucchiara. Boosting modern and historical handwrit- ten text recognition with deformable convolutions. Interna- tional Journal on Document Analysis and Recognition (IJ- DAR), 25(3):207–217, 2022. 6

    work page 2022

Show all 59 references
  1. [7]

    Kartik Chaudhary and Raghav Bali. Easter2. 0: Improving convolutional models for handwritten text recognition.arXiv preprint arXiv:2205.14879, 2022. 6

    2022 arXiv

  2. [8]

    Contrastive test-time adaptation

    Dian Chen, Dequan Wang, Trevor Darrell, and Sayna Ebrahimi. Contrastive test-time adaptation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pat- tern Recognition, pages 295–305, 2022. 3

    2022

  3. [9]

    Improved test-time adaptation for domain generalization

    Liang Chen, Yong Zhang, Yibing Song, Ying Shan, and Lingqiao Liu. Improved test-time adaptation for domain generalization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 24172– 24182, 2023. 3

    2023

  4. [10]

    Metafscil: A meta-learning approach for few-shot class incremental learning

    Zhixiang Chi, Li Gu, Huan Liu, Yang Wang, Yuanhao Yu, and Jin Tang. Metafscil: A meta-learning approach for few-shot class incremental learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 14166–14175, 2022. 2

    2022

  5. [11]

    Adapting to distribution shift by visual domain prompt generation

    Zhixiang Chi, Li Gu, Tao Zhong, Huan Liu, Yuanhao Yu, Konstantinos N Plataniotis, and Yang Wang. Adapting to distribution shift by visual domain prompt generation. arXiv preprint arXiv:2405.02797, 2024. 2

    2024 arXiv

  6. [12]

    Test- time fast adaptation for dynamic scene deblurring via meta- auxiliary learning

    Zhixiang Chi, Yang Wang, Yuanhao Yu, and Jin Tang. Test- time fast adaptation for dynamic scene deblurring via meta- auxiliary learning. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 9137–9146, 2021. 2, 3

    2021

  7. [13]

    Span: a simple predict & align network for handwritten para- graph recognition

    Denis Coquenet, Cl ´ement Chatelain, and Thierry Paquet. Span: a simple predict & align network for handwritten para- graph recognition. In International Conference on Document Analysis and Recognition, pages 70–84. Springer, 2021. 6

    2021

  8. [14]

    End-to-end handwritten paragraph text recognition using a vertical attention network

    Denis Coquenet, Cl ´ement Chatelain, and Thierry Paquet. End-to-end handwritten paragraph text recognition using a vertical attention network. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(1):508–524, 2022. 1, 6

    2022

  9. [15]

    Dan: a segmentation-free document attention network for handwritten document recognition

    Denis Coquenet, Cl ´ement Chatelain, and Thierry Paquet. Dan: a segmentation-free document attention network for handwritten document recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023. 1, 3, 5, 6, 7, 8

    2023

  10. [16]

    Faster dan: Multi-target queries with document positional encoding for end-to-end handwritten document recognition

    Denis Coquenet, Cl ´ement Chatelain, and Thierry Paquet. Faster dan: Multi-target queries with document positional encoding for end-to-end handwritten document recognition. In International Conference on Document Analysis and Recognition, pages 182–199. Springer, 2023. 6, 7

    2023

  11. [17]

    Bert: Pre-training of deep bidirectional transformers for language understanding

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805, 2018. 3

    2018 arXiv

  12. [19]

    The pascal visual object classes (voc) challenge

    Mark Everingham, Luc Van Gool, Christopher KI Williams, John Winn, and Andrew Zisserman. The pascal visual object classes (voc) challenge. International Journal of Computer Vision, 88:303–338, 2010. 7

    2010

  13. [20]

    Model- agnostic meta-learning for fast adaptation of deep networks

    Chelsea Finn, Pieter Abbeel, and Sergey Levine. Model- agnostic meta-learning for fast adaptation of deep networks. Proceedings of the 34th International Conference on Ma- chine Learning, 70:1126–1135, 2017. 2, 4

    2017

  14. [21]

    Convolutional neural network based intelligent handwritten document recognition

    TM Ghazal. Convolutional neural network based intelligent handwritten document recognition. Computers, Materials & Continua, 70(3):4563–4581, 2022. 1

    2022

  15. [22]

    Handwritten gujarati numer- als classification based on deep convolution neural networks using transfer learning scenarios

    Parth Goel and Amit Ganatra. Handwritten gujarati numer- als classification based on deep convolution neural networks using transfer learning scenarios. IEEE Access, 11:20202– 20215, 2023. 5

    2023

  16. [23]

    Icdar 2011-french handwriting recognition competition

    Emmanuele Grosicki and Haikal El-Abed. Icdar 2011-french handwriting recognition competition. In 2011 International Conference on Document Analysis and Recognition , pages 1459–1463. IEEE, 2011. 6, 7

    2011

  17. [24]

    Improving protonet for few-shot video object recog- nition: Winner of orbit challenge 2022

    Li Gu, Zhixiang Chi, Huan Liu, Yuanhao Yu, and Yang Wang. Improving protonet for few-shot video object recog- nition: Winner of orbit challenge 2022. arXiv preprint arXiv:2210.00174, 2022. 2

    2022 arXiv

  18. [25]

    Masked autoencoders are scalable vision learners

    Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Doll´ar, and Ross Girshick. Masked autoencoders are scalable vision learners. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 16000– 16009, 2022. 3, 4

    2022

  19. [26]

    General models for handwritten text recognition: Feasibility and state-of-the art

    Tobias Mathias Hodel, David Selim Schoch, Christa Schnei- der, and Jake Purcell. General models for handwritten text recognition: Feasibility and state-of-the art. german kurrent as an example. Journal of Open Humanities Data, 7(13):1– 10, 2021. 1

    2021

  20. [27]

    Johnson and Emily D

    Michael C. Johnson and Emily D. Richards. Sequential handwriting analysis with lstm networks. Pattern Recogni- tion Letters, 142:45–51, 2021. 2

    2021

  21. [28]

    Pay attention to what you read: Non-recurrent handwritten text- line recognition (2020)

    L Kang, P Riba, M Rusi ˜nol, A Forn´es, and M Villegas. Pay attention to what you read: Non-recurrent handwritten text- line recognition (2020). arXiv preprint arXiv:2005.13044 ,

    2020 arXiv

  22. [29]

    Meta- sgd: Learning to learn quickly for few-shot learning

    Zhenguo Li, Fengwei Zhou, Fei Chen, and Hang Li. Meta- sgd: Learning to learn quickly for few-shot learning. In Pro- ceedings of the 31st International Conference on Neural In- formation Processing Systems, pages 3637–3647, 2017. 2

    2017

  23. [30]

    Self-supervised spa- tiotemporal representation learning by exploiting video con- tinuity

    Hanwen Liang, Niamul Quader, Zhixiang Chi, Lizhe Chen, Peng Dai, Juwei Lu, and Yang Wang. Self-supervised spa- tiotemporal representation learning by exploiting video con- tinuity. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, pages 1564–1573, 2022. 3

    2022

  24. [31]

    Microsoft coco: Common objects in context

    Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll´ar, and C Lawrence Zitnick. Microsoft coco: Common objects in context. In Computer Vision–ECCV 2014: 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proceeding...

    2014

  25. [32]

    Meta-auxiliary learning for future depth pre- diction in videos

    Huan Liu, Zhixiang Chi, Yuanhao Yu, Yang Wang, Jun Chen, and Jin Tang. Meta-auxiliary learning for future depth pre- diction in videos. In Proceedings of the IEEE/CVF Win- ter Conference on Applications of Computer Vision , pages 5756–5765, 2023. 2

    2023

  26. [33]

    Few-shot class-incremental learning via entropy-regularized data-free replay

    Huan Liu, Li Gu, Zhixiang Chi, Yang Wang, Yuanhao Yu, Jun Chen, and Jin Tang. Few-shot class-incremental learning via entropy-regularized data-free replay. In European Con- ference on Computer Vision, pages 146–162. Springer, 2022. 2

    2022

  27. [34]

    Maskocr: Text recognition with masked encoder-decoder pretraining

    Pengyuan Lyu, Chengquan Zhang, Shanshan Liu, Meina Qiao, Yangliu Xu, Liang Wu, Kun Yao, Junyu Han, Er- rui Ding, and Jingdong Wang. Maskocr: Text recognition with masked encoder-decoder pretraining. arXiv preprint arXiv:2206.00311, 2022. 3

    2022 arXiv

  28. [35]

    Crowd counting us- ing meta-test-time adaptation

    Chaoqun Ma, Ferrante Neri, Li Gu, Ziqiang Wang, Jian Wang, Anyong Qing, and Yang Wang. Crowd counting us- ing meta-test-time adaptation. International journal of neu- ral systems, 34(11):2450061, 2024. 2

    2024

  29. [36]

    The iam-database: an english sentence database for offline handwriting recognition

    U-V Marti and Horst Bunke. The iam-database: an english sentence database for offline handwriting recognition. In- ternational journal on document analysis and recognition , 5:39–46, 2002. 6, 7

    2002

  30. [37]

    On first-order meta-learning algorithms

    Alex Nichol, Joshua Achiam, and John Schulman. On first-order meta-learning algorithms. arXiv preprint arXiv:1803.02999, 2018. 2

    2018 arXiv

  31. [38]

    Handwritten kazakh and russian (hkr) database for text recognition

    Daniyar Nurseitov, Kairat Bostanbekov, Daniyar Kur- mankhojayev, Anel Alimova, Abdelrahman Abdallah, and Rassul Tolegenov. Handwritten kazakh and russian (hkr) database for text recognition. Multimedia Tools and Appli- cations, 80(21):33075–33097, 2021. 1

    2021

  32. [39]

    Meta-learning of pooling layers for character recognition

    Takato Otsuzuki, Heon Song, Seiichi Uchida, and Hideaki Hayashi. Meta-learning of pooling layers for character recognition. In Document Analysis and Recognition–ICDAR 2021: 16th International Conference, Lausanne, Switzer- land, September 5–10, 2021, Proceedings, Part III 16, ...

    2021

  33. [40]

    Joan Puigcerver. Are multidimensional recurrent layers re- ally necessary for handwritten text recognition? In 2017 14th IAPR International Conference on Document Analysis and Recognition (ICDAR) , volume 1, pages 67–72. IEEE,

    2017

  34. [41]

    Meta self- learning for multi-source domain adaptation: a benchmark

    Shuhao Qiu, Chuang Zhu, and Wenli Zhou. Meta self- learning for multi-source domain adaptation: a benchmark. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 1592–1601, 2021. 2

    2021

  35. [42]

    Pat- tern recognition-recognition of handwritten document using convolutional neural networks

    M Rajalakshmi, P Saranya, and P Shanmugavadivu. Pat- tern recognition-recognition of handwritten document using convolutional neural networks. In 2019 IEEE International Conference on Intelligent Techniques in Control, Optimiza- tion and Signal Processing (INCOS), pages 1–7. I...

    2019

  36. [43]

    Transformer-based approach for joint handwriting and named entity recognition in historical doc- ument

    Ahmed Cheikh Rouhou, Marwa Dhiaf, Yousri Kessentini, and Sinda Ben Salem. Transformer-based approach for joint handwriting and named entity recognition in historical doc- ument. Pattern Recognition Letters, 155:128–134, 2022. 6

    2022

  37. [44]

    Icfhr2016 competition on hand- written text recognition on the read dataset

    Joan Andreu Sanchez, Ver ´onica Romero, Alejandro H Toselli, and Enrique Vidal. Icfhr2016 competition on hand- written text recognition on the read dataset. In 2016 15th In- ternational Conference on Frontiers in Handwriting Recog- nition (ICFHR), pages 630–635. IEEE, 2016. 6, 7

    2016

  38. [45]

    Image quality assessment through fsim, ssim, mse and psnr—a comparative study

    Umme Sara, Morium Akter, and Mohammad Shorif Ud- din. Image quality assessment through fsim, ssim, mse and psnr—a comparative study. Journal of Computer and Com- munications, 7(3):8–18, 2019. 5

    2019

  39. [46]

    Psnr vs ssim: impercepti- bility quality assessment for image steganography

    De Rosal Igantius Moses Setiadi. Psnr vs ssim: impercepti- bility quality assessment for image steganography. Multime- dia Tools and Applications, 80(6):8423–8444, 2021. 5

    2021

  40. [47]

    Full page handwriting recognition via image to sequence extraction

    Sumeet S Singh and Sergey Karayev. Full page handwriting recognition via image to sequence extraction. In Document Analysis and Recognition–ICDAR 2021: 16th International Conference, Lausanne, Switzerland, September 5–10, 2021, Proceedings, Part III 16, pages 55–69. Springer, 2021. 6

    2021

  41. [48]

    Test-time training with self- supervision for generalization under distribution shifts

    Yu Sun, Xiaolong Wang, Zhuang Liu, John Miller, Alexei Efros, and Moritz Hardt. Test-time training with self- supervision for generalization under distribution shifts. InIn- ternational Conference on Machine Learning , pages 9229–

  42. [49]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszko- reit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in Neural Information Processing Systems, 30, 2017. 3

    2017

  43. [50]

    Tent: Fully test-time adaptation by entropy minimization

    Dequan Wang, Evan Shelhamer, Shaoteng Liu, Bruno Ol- shausen, and Trevor Darrell. Tent: Fully test-time adaptation by entropy minimization. arXiv preprint arXiv:2006.10726,

    2006 arXiv

  44. [51]

    Combinatorial learning of graph edit distance via dynamic embedding

    Runzhong Wang, Tianqi Zhang, Tianshu Yu, Junchi Yan, and Xiaokang Yang. Combinatorial learning of graph edit distance via dynamic embedding. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5241–5250, 2021. 7

    2021

  45. [52]

    Decoupled attention network for text recognition

    Tianwei Wang, Yuanzhi Zhu, Lianwen Jin, Canjie Luo, Xi- aoxue Chen, Yaqiang Wu, Qianying Wang, and Mingxi- ang Cai. Decoupled attention network for text recognition. In Proceedings of the AAAI conference on artificial intelli- gence, volume 34, pages 12216–12224, 2020. 6

    2020

  46. [53]

    Distribution align- ment for fully test-time adaptation with dynamic online data streams

    Ziqiang Wang, Zhixiang Chi, Yanan Wu, Li Gu, Zhi Liu, Konstantinos Plataniotis, and Yang Wang. Distribution align- ment for fully test-time adaptation with dynamic online data streams. In European Conference on Computer Vision, pages 332–349. Springer, 2025. 3

    2025

  47. [54]

    Start, follow, read: End-to-end full-page handwriting recognition

    Curtis Wigington, Chris Tensmeyer, Brian Davis, William Barrett, Brian Price, and Scott Cohen. Start, follow, read: End-to-end full-page handwriting recognition. In Pro- ceedings of the European Conference on Computer Vision (ECCV), pages 367–383, 2018. 6

    2018

  48. [56]

    Metagcd: Learning to continually learn in generalized cat- egory discovery

    Yanan Wu, Zhixiang Chi, Yang Wang, and Songhe Feng. Metagcd: Learning to continually learn in generalized cat- egory discovery. In Proceedings of the IEEE/CVF Interna- tional Conference on Computer Vision , pages 1655–1665,

  49. [57]

    Test-time domain adaptation by learning domain-aware batch normalization

    Yanan Wu, Zhixiang Chi, Yang Wang, Konstantinos N Pla- taniotis, and Songhe Feng. Test-time domain adaptation by learning domain-aware batch normalization. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 15961–15969, 2024. 2

    2024

  50. [58]

    A-vit: Adaptive to- kens for efficient vision transformer

    Hongxu Yin, Arash Vahdat, Jose M Alvarez, Arun Mallya, Jan Kautz, and Pavlo Molchanov. A-vit: Adaptive to- kens for efficient vision transformer. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10809–10818, 2022. 3

    2022

  51. [59]

    Meta-dmoe: Adapting to domain shift by meta- distillation from mixture-of-experts

    Tao Zhong, Zhixiang Chi, Li Gu, Yang Wang, Yuanhao Yu, and Jin Tang. Meta-dmoe: Adapting to domain shift by meta- distillation from mixture-of-experts. Advances in Neural In- formation Processing Systems, 35:22243–22257, 2022. 2

    2022

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.