Pith. sign in

REVIEW 2 major objections 5 minor 69 references

Online Writer Retrieval with Chinese Handwritten Phrases: A Synergistic Temporal-Frequency Representation Learning Approach

T0 review · 2 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read This paper claims that online writer retrieval on short Chinese handwritten phrases improves sharply when the model learns temporal and frequency representations together, and it introduces the DOLPHIN model and the OLIWER dataset to back…

desk verdict A genuinely useful benchmark and strong baseline for online writer retrieval, but the head-to-head gains need a careful look because the 2D-to-1D baseline adaptations may have weakened the competition. read the letter →

arxiv 2412.11668 v2 pith:SQ5CB6LE submitted 2024-12-16 cs.CV

Peirong Zhang , Lianwen Jin This is my paper · ORCID
classification cs.CV
keywords onlinewriterretrievalhandwrittenphrasetemporal-frequencyrepresentationdiscretewavelettransformgatedcross-attentionchannelactivationOLIWERdatasetChinesehandwriting
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

This paper tries to establish that online writer retrieval—finding all handwriting samples by the same person from short online handwritten phrases—can be made substantially more accurate by jointly learning temporal and frequency representations, and that the field has lacked a large-scale benchmark to measure this. To that end, it introduces DOLPHIN, a 1D CNN retrieval model, and OLIWER, a dataset of 674,017 Chinese handwritten phrases from 1,731 writers, and reports that DOLPHIN improves Rank-1 accuracy to 96.40% and mAP to 69.83% on OLIWER, against 93.32% and 61.01% for the best prior method. The paper also claims that cross-domain retrieval performance depends critically on aligning point sampling frequency and pressure information between datasets. A sympathetic reader would care because reliable phrase-level writer retrieval would support forensic search and handwriting analysis on the short, dynamic samples that dominate real online writing.

What carries the argument

The load-bearing object is the HFGA block, a gated cross-attention module that applies a 1D discrete wavelet transform to split an input handwriting sequence into high-frequency and low-frequency sub-bands, then uses the high-frequency band as keys and values and the original temporal sequence as queries, with a tanh-gated residual connection. It lets the model amplify stroke-level details that are hard to see in the time domain alone. The CAIR block, a channel-split inverted residual with squeeze-excitation and channel shuffle, is the temporal counterpart that reduces channel redundancy, and a Context-Aware FPN adds multi-scale feature fusion with global context. Together these modules form DOLPHIN, a 1D CNN trained with Circle loss, OIM loss, and label-smoothed writer-ID loss, and evaluated by cosine-similarity ranking.

What would settle it

Inspect the 5,100 added DCOH-E lines: if the claimed 255 additional writers do not appear in the cited DCOH release or any named source, and if Algorithm 1's stroke-interval threshold, when compared against manual character boundaries on a random sample, frequently assigns strokes to wrong characters, then the OLIWER labels and every comparison built on them would be corrupted.

Watch

Extended reading notes

Core claim

The central claim is that a representation which combines the original temporal trace of an online handwriting sample with its high-frequency sub-bands, extracted via a discrete wavelet transform, captures more of a writer's individual style than temporal-only or existing retrieval representations. DOLPHIN operationalizes this with the HFGA block, which runs gated cross-attention between the temporal sequence and the high-frequency components so that local writing details such as stroke curvature and pressure variation are amplified, while the CAIR block reduces channel redundancy in the temporal backbone. Reports show DOLPHIN outperforming all compared methods on four datasets (OLIWER, segmented CASIA-OLHWDB2, segmented DCOH-E, and SCUT-COUCH2009), with the largest gains on mAP; ablation studies attribute the gains to CAIR, HFGA, and Context-Aware FPN, and cross-domain experiments support the claim that raising sampling frequency from 30Hz to 120Hz and adding pressure information narrows the distribution gap between handwriting domains.

Load-bearing premise

The load-bearing premise is that OLIWER is a valid collection of correctly labeled handwriting from 1,731 distinct writers, which requires that the 255 additional writers added to DCOH to form DCOH-E come from a traceable source and that Algorithm 1's stroke-interval threshold cuts text lines at true character boundaries.

Editorial extensions

If this is right

  • DOLPHIN sets a new state of the art for open-set online writer retrieval on phrase-level Chinese handwriting, including on mAP, where it surpasses the best prior method by 8.82 points on OLIWER.
  • OLIWER provides a community benchmark of 674,017 phrases from 1,731 writers, filling the large-scale dataset gap the paper identifies in online writer retrieval.
  • Because all models are evaluated with the same 14 hand-crafted time functions, the reported margins are attributable to architecture rather than input preprocessing.
  • Cross-domain results imply that future online handwriting datasets should record pressure and use a 120Hz sampling rate to make retrieval systems transfer across domains.
  • DOLPHIN's 2.14M parameters and roughly 9.95ms per-sample inference time make it plausible for real-time forensic search on large galleries.

Reading between the lines

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

  • The OLIWER dataset is an aggregation of existing sources rather than a newly collected corpus, so its utility depends on the integrity of writer identity and phrase segmentation in the source datasets; an independent audit of that lineage would be the strongest validation of the benchmark.
  • The conclusion that sampling frequency and pressure are the dominant domain-gap factors is supported by controlled experiments, but the direction of causality could be tested further by training on native 120Hz data with pressure artificially removed without resampling artifacts.
  • The architecture is language-agnostic in principle, but this paper only tests Chinese data; applying the same pipeline to English, Arabic, or Vietnamese online handwriting after phrase-level segmentation would test whether the temporal-frequency synergy transfers.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 5 minor

Summary. The paper proposes DOLPHIN, a 1D convolutional model for online writer retrieval from Chinese handwritten phrases, combining temporal and frequency features through two new blocks (HFGA for gated cross-attention over high-frequency DWT sub-bands, and CAIR for channel activation), plus a Context-Aware FPN. It also introduces OLIWER, a large aggregation dataset built from CASIA-OLHWDB, DCOH-E, and SCUT-COUCH2009. Experiments on OLIWER, segmented CASIA-OLHWDB2, segmented DCOH-E, and SCUT-COUCH2009 report large improvements over adapted baselines, with ablations, efficiency comparisons, writer identification results, and cross-domain analyses on sampling frequency and pressure.

Significance. If the results hold, the paper makes two substantive contributions: a new large-scale benchmark (OLIWER) for a relatively underexplored task, and a strong retrieval model with clear ablations showing the value of each proposed component. The empirical work is extensive: four datasets, repeated trials with reported standard deviations, module ablations, efficiency measurements, and additional closed-set identification and cross-domain experiments. The promised release of code and dataset is a further practical strength. However, the headline claim of state-of-the-art performance depends on two load-bearing points that are not fully established: the fairness and completeness of the 2D-to-1D baseline adaptations, and the traceability of the OLIWER/DCOH-E construction, especially the provenance of 255 additional DCOH-E writers and the validation of the stroke-based phrase segmentation. The cross-domain conclusions are also mixed: DOLPHIN underperforms Sig2Vec in one transfer direction, which the authors acknowledge.

major comments (2)
  1. [§V-E, Table III] The central comparison rests on the claim that 2D baselines were converted to 1D 'by simply substituting the Conv2d module with the Conv1d module in Pytorch implementations.' This is not a well-defined operation for the strongest baselines: CAL uses counterfactual attention over 2D spatial feature maps, OSNet relies on omni-scale 2D convolutions with multi-scale context aggregation, CDNet uses 2D group convolutions in a combined depth space, and EfficientNet/MobileNetV2 use SE blocks with 2D global average pooling. A literal Conv2d-to-Conv1d substitution would either fail to run (e.g., BatchNorm2d with 3D inputs) or silently disable these central mechanisms. If the authors instead converted all associated 2D operations (normalization, pooling, attention), that conversion is not 'simple' and needs to be documented and analyzed. Without a specification of the exact mapping and ideally released code for the adapted baselines, the large margins in Table III may be inflated by systematically weakened baselines. Please provide the adapted implementations and a per-baseline description of how attention/pooling/SE mechanisms were mapped to 1D, and confirm that each baseline's core design was preserved.
  2. [§IV, DCOH-E/OLIWER] The construction of OLIWER is load-bearing for every reported comparison, and two parts are not adequately supported. First, Section IV states that DCOH-E is formed by adding 5,100 lines from 255 additional writers to the Chinese subset of DCOH, but it gives no source, citation, or collection details for these additional writers. Without provenance, the writer identities and sample counts in OLIWER and DCOH-E cannot be verified, and any corrupted identities would directly affect all retrieval labels. Second, the phrase segmentation in Algorithm 1 uses as threshold the time interval ranked at position N_Y+1; this heuristic is never validated against ground-truth character boundaries. Since CASIA-OLHWDB has character-level annotations, a comparison of Algorithm 1's output against those annotations would provide a quantitative check, and a manual or automated validation for DCOH-E should be reported. Please document the origin and licensing of the additional writer data, confirm that no writer overlaps across the three constituent datasets, and validate the segmentation heuristic.
minor comments (5)
  1. [§V-E vs Table III caption] There is an inconsistency in the reported number of repetitions: Section V-E says 'we repeated each experiment 30 times,' while the Table III caption says 'Each experiment is repeated 50 times.' Please correct this discrepancy and ensure the stated number matches the actual experiments.
  2. [Eq. (2)-(3), §III-B] The attention computation is written as w = k^T @ q, with k in R^{⌊L/2⌋×d} and q in R^{L×d}. With @ denoting matrix multiplication, k^T @ q is undefined because the inner dimensions do not match; the intended expression is presumably q @ k^T (or an equivalent transposed formulation), and the softmax dimension should be described consistently with that expression. Please correct the equation to make the HFGA block reproducible.
  3. [§V-I and Conclusion] The cross-domain results are mixed: DOLPHIN underperforms Sig2Vec on OLHWDB2→DCOH-E (57.02% vs 62.28% Rank-1; 21.26% vs 24.55% mAP), while outperforming others in the reverse direction. This is acknowledged in the text, but the conclusion and abstract should be worded to avoid implying uniform superiority in cross-domain settings; a short discussion of why the proposed model is less transferable in that direction would be helpful.
  4. [Table IV] The ablation table is difficult to parse: several rows share the same combination of checkmarks in the main text (e.g., the rows referenced as 'line 5' and 'line 7' are not unambiguously labeled), and the textual references to line numbers do not align with a clear row enumeration. Please add explicit row labels or a legend so that each configuration can be identified.
  5. [§IV, DCOH-E bullet] The text says the segmentation details are 'included in Procedure 1 in supplementary files,' but Algorithm 1 appears in the main text immediately afterward. Please reconcile this cross-reference so readers know where the algorithm is defined.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found; DOLPHIN's superiority is an empirical, externally benchmarked result, and the noted baseline-conversion and dataset-provenance issues are validity risks rather than circular derivations.

full rationale

This is an empirical systems and dataset paper rather than a formal derivation chain. The central claims—DOLPHIN outperforms prior methods and OLIWER is a large-scale online writer retrieval dataset—are established by training on fixed writer-disjoint splits and comparing against external baselines, not by algebraically reducing a target quantity to an input. No parameter is fitted to the evaluation set and then renamed as a prediction: the model hyperparameters and losses are fixed before evaluation and applied uniformly to all compared methods. The ablations add or remove DOLPHIN components on the same fixed protocol, which is standard empirical attribution and does not make the result self-definitional. The paper's self-citations (e.g., SCUT-COUCH2009, Sig2Vec, MSDS) provide datasets, components, or preprocessing conventions, but none is invoked as the proof of DOLPHIN's superiority. The cross-domain observations about sampling frequency and pressure are post hoc empirical analyses that explain the already-adopted preprocessing choices; they are not fitted targets, and the paper does not present them as predictions derived from the model. The undocumented 255 additional DCOH-E writers and the 'simply substituting Conv2d with Conv1d' baseline adaptation are legitimate threats to dataset traceability and comparison fairness, but they do not exhibit a claimed result reducing to its own input by construction. No circular step can be quoted from the equations or the evaluation protocol, so the appropriate finding is no significant circularity.

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

The central claim is empirical and does not derive a numeric constant, but the benchmark and model rest on several unvalidated assumptions about the data and the features.

free parameters (5)
  • Circle loss margin m = 0.25
    Hyperparameter in Eq. 16; set by hand, affects the representation learning and retrieval results.
  • Circle loss scale gamma = 32
    Hyperparameter in Eq. 14; chosen without reported tuning.
  • Label smoothing epsilon = 0.1
    Used in L_ID; affects classification loss in open-set training.
  • Resampling frequency = 120 Hz
    All sequences resampled to 120 Hz; the choice is justified post hoc by the cross-domain experiment in Section V-I.
  • Phrase length range = 2 to 5 characters
    Dataset construction choice in Section IV; may affect retrieval difficulty.
assumptions (3)
  • domain assumption The high-frequency components of the 1D discrete wavelet transform carry writer-specific discriminative details.
    Invoked in Section III-B, HFGA block; if false, the frequency branch adds noise rather than signal.
  • domain assumption The 14 time functions derived from signature verification (Table II) are appropriate inputs for online Chinese writer retrieval.
    Section V-C; all models use these features; if they omit crucial information, results are conditioned on that choice.
  • domain assumption The three public datasets, after segmentation and standardization, form a coherent benchmark without spurious writer overlap.
    Section IV; writer identities are taken from the original datasets, and DCOH-E adds 255 writers whose origin is not documented.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Online Writer Retrieval with Chinese Handwritten Phrases: A Synergistic Temporal-Frequency Representation Learning Approach." pith.science (2026). https://pith.science/paper/SQ5CB6LE

@misc{pith2026241211668,
  author       = {Pith},
  title        = {Pith review of: Online Writer Retrieval with Chinese Handwritten Phrases: A Synergistic Temporal-Frequency Representation Learning Approach},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SQ5CB6LE}},
  note         = {Machine review of arXiv:2412.11668}
}
read the original abstract

Currently, the prevalence of online handwriting has spurred a critical need for effective retrieval systems to accurately search relevant handwriting instances from specific writers, known as online writer retrieval. Despite the growing demand, this field suffers from a scarcity of well-established methodologies and public large-scale datasets. This paper tackles these challenges with a focus on Chinese handwritten phrases. First, we propose DOLPHIN, a novel retrieval model designed to enhance handwriting representations through synergistic temporal-frequency analysis. For frequency feature learning, we propose the HFGA block, which performs gated cross-attention between the vanilla temporal handwriting sequence and its high-frequency sub-bands to amplify salient writing details. For temporal feature learning, we propose the CAIR block, tailored to promote channel interaction and reduce channel redundancy. Second, to address data deficit, we introduce OLIWER, a large-scale online writer retrieval dataset encompassing over 670,000 Chinese handwritten phrases from 1,731 individuals. Through extensive evaluations, we demonstrate the superior performance of DOLPHIN over existing methods. In addition, we explore cross-domain writer retrieval and reveal the pivotal role of increasing feature alignment in bridging the distributional gap between different handwriting data. Our findings emphasize the significance of point sampling frequency and pressure features in improving handwriting representation quality and retrieval performance. Code and dataset are available at https://github.com/SCUT-DLVCLab/DOLPHIN.

Figures

Figures reproduced from arXiv: 2412.11668 by the authors.

Figure 1
Figure 1. Comparison between writer identification and writer re [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The overall architecture of the proposed writer retrieval model DOLPHIN. DOLPHIN is a 1D CNN-based model, [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 4
Figure 4. Schematic of the HFGA block. informative channel features. Finally, we pass x ′ through a CBA module and a dropout operation to get the output features. The channel split-concat scheme, 1D IR, SE module, and channel shuffle operation progressively activate the informa￾tion flow across the channel dimension, effectively diminishing the redundancy of channel-wise features and thus empowering channel modeling. For the … view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: Schematic of the Context-Aware FPN. Transform (DWT) to decompose x t into the high frequency components x f h ∈ R ⌊L/2⌋×d and low frequency components x f l ∈ R ⌊L/2⌋×d . Here, x f h corresponds to the sharp turns and fine-grained details of the signal, which represent…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

69 extracted references · 67 canonical work pages

  1. [1]

    Writer Retrieval - Exploration of a Novel Biometric Scenario using Perceptual Features Derived from Script Orientation,

    V . Atanasiu, L. Likforman-Sulem, and N. Vincent, “Writer Retrieval - Exploration of a Novel Biometric Scenario using Perceptual Features Derived from Script Orientation,” in ICDAR, 2011, pp. 628–632

    work page 2011

  2. [2]

    Writer Identi- fication using Machine Learning Approaches: A Com- prehensive Review,

    A. Rehman, S. Naz, and M. I. Razzak, “Writer Identi- fication using Machine Learning Approaches: A Com- prehensive Review,” Multimedia Tools and Applications , vol. 78, pp. 10 889–10 931, 2019

    work page 2019

  3. [3]

    Writer Retrieval and Writer Identification using Local Features,

    S. Fiel and R. Sablatnig, “Writer Retrieval and Writer Identification using Local Features,” in DAS, 2012, pp. 145–149

    work page 2012

  4. [4]

    Learning Fea- tures for Writer Retrieval and Identification using Triplet CNNs,

    M. Keglevic, S. Fiel, and R. Sablatnig, “Learning Fea- tures for Writer Retrieval and Identification using Triplet CNNs,” in ICFHR, 2018, pp. 211–216

    work page 2018

  5. [5]

    DeepWriterID: An End-to- End Online Text-Independent Writer Identification Sys- tem,

    W. Yang, L. Jin, and M. Liu, “DeepWriterID: An End-to- End Online Text-Independent Writer Identification Sys- tem,” IEEE Intelligent Systems, vol. 31, no. 2, pp. 45–53, 2016

    work page 2016

  6. [6]

    Level Online Writer Identification,

    Z. Chen, H.-X. Yu, A. Wu, and W.-S. Zheng, “Level Online Writer Identification,” International Journal of Computer Vision, vol. 129, no. 5, pp. 1394–1409, 2021

    work page 2021

  7. [7]

    Flamingo: A Visual Language Model for Few-Shot Learning,

    J.-B. Alayrac, J. Donahue, P. Luc, A. Miech, I. Barr, Y . Hasson, K. Lenc, A. Mensch, K. Millican, M. Reynolds, R. Ring, E. Rutherford, S. Cabi, T. Han, Z. Gong, S. Samangooei, M. Monteiro, J. L. Menick, IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY , VOL. 19, 2024 13 S. Borgeaud, A. Brock, A. Nematzadeh, S. Sharifzadeh, M. a. Bi ´nkowski, R. Ba...

    work page 2024

  8. [8]

    MobileNetV2: Inverted Residuals and Linear Bottlenecks,

    M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L.- C. Chen, “MobileNetV2: Inverted Residuals and Linear Bottlenecks,” in CVPR, 2018

    work page 2018

Show all 69 references
  1. [9]

    ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices,

    X. Zhang, X. Zhou, M. Lin, and J. Sun, “ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices,” in CVPR, 2018

    2018

  2. [10]

    MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications,

    A. G. Howard, M. Zhu, B. Chen, D. Kalenichenko, W. Wang, T. Weyand, M. Andreetto, and H. Adam, “MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications,” arXiv preprint arXiv:1704.04861, 2017

    2017 arXiv

  3. [11]

    Feature Pyramid Networks for Object Detection,

    T.-Y . Lin, P. Dollar, R. Girshick, K. He, B. Hariharan, and S. Belongie, “Feature Pyramid Networks for Object Detection,” in CVPR, 2017

    2017

  4. [12]

    GCNet: Non- Local Networks Meet Squeeze-Excitation Networks and Beyond,

    Y . Cao, J. Xu, S. Lin, F. Wei, and H. Hu, “GCNet: Non- Local Networks Meet Squeeze-Excitation Networks and Beyond,” in ICCVW, 2019

    2019

  5. [13]

    CASIA Online and Offline Chinese Handwriting Databases,

    C.-L. Liu, F. Yin, D.-H. Wang, and Q.-F. Wang, “CASIA Online and Offline Chinese Handwriting Databases,” in ICDAR, 2011, pp. 37–41

    2011

  6. [14]

    EGO-LM: An Efficient, Generic, and Out-of-the-box Language Model for Hand- written Text Recognition,

    H. Li, D. Peng, and L. Jin, “EGO-LM: An Efficient, Generic, and Out-of-the-box Language Model for Hand- written Text Recognition,” Pattern Recognition, 2025

    2025

  7. [15]

    SCUT- COUCH2009—A Comprehensive Online Unconstrained Chinese Handwriting Database and Benchmark Evalua- tion,

    L. Jin, Y . Gao, G. Liu, Y . Li, and K. Ding, “SCUT- COUCH2009—A Comprehensive Online Unconstrained Chinese Handwriting Database and Benchmark Evalua- tion,” International Journal on Document Analysis and Recognition, vol. 14, pp. 53–64, 2011

    2011

  8. [16]

    GR-RNN: Global-Context Residual Recurrent Neural Networks for Writer Identifi- cation,

    S. He and L. Schomaker, “GR-RNN: Global-Context Residual Recurrent Neural Networks for Writer Identifi- cation,” Pattern Recognition, vol. 117, p. 107975, 2021

    2021

  9. [17]

    SynSig2Vec: Forgery-Free Learning of Dynamic Signature Represen- tations by Sigma Lognormal-Based Synthesis and 1D CNN,

    S. Lai, L. Jin, Y . Zhu, Z. Li, and L. Lin, “SynSig2Vec: Forgery-Free Learning of Dynamic Signature Represen- tations by Sigma Lognormal-Based Synthesis and 1D CNN,” IEEE Transactions on Pattern Analysis and Ma- chine Intelligence, vol. 44, no. 10, pp. 6472–6485, 2022

    2022

  10. [18]

    Combined Depth Space Based Architecture Search for Person Re-Identification,

    H. Li, G. Wu, and W.-S. Zheng, “Combined Depth Space Based Architecture Search for Person Re-Identification,” in CVPR, 2021, pp. 6729–6738

    2021

  11. [19]

    Counterfactual At- tention Learning for Fine-Grained Visual Categorization and Re-Identification,

    Y . Rao, G. Chen, J. Lu, and J. Zhou, “Counterfactual At- tention Learning for Fine-Grained Visual Categorization and Re-Identification,” in ICCV, 2021, pp. 1025–1034

    2021

  12. [20]

    Learning Generalisable Omni-Scale Representations for Person Re-Identification,

    K. Zhou, Y . Yang, A. Cavallaro, and T. Xiang, “Learning Generalisable Omni-Scale Representations for Person Re-Identification,” IEEE Transactions on Pattern Anal- ysis and Machine Intelligence , vol. 44, no. 9, pp. 5056– 5069, 2022

    2022

  13. [21]

    Aggregating Local Image Descriptors into Compact Codes,

    H. J ´egou, F. Perronnin, M. Douze, J. S ´anchez, P. P ´erez, and C. Schmid, “Aggregating Local Image Descriptors into Compact Codes,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 34, no. 9, pp. 1704–1716, 2012

    2012

  14. [22]

    Writer Identification and Writer Retrieval using the Fisher Vector on Visual V ocabularies,

    S. Fiel and R. Sablatnig, “Writer Identification and Writer Retrieval using the Fisher Vector on Visual V ocabularies,” in ICDAR, 2013, pp. 545–549

    2013

  15. [23]

    Unsu- pervised Feature Learning for Writer Identification and Writer Retrieval,

    V . Christlein, M. Gropp, S. Fiel, and A. Maier, “Unsu- pervised Feature Learning for Writer Identification and Writer Retrieval,” in ICDAR, 2017, pp. 991–997

    2017

  16. [24]

    Writer Identification and Writer Retrieval Based on NetVLAD with Re- Ranking,

    S. Rasoulzadeh and B. BabaAli, “Writer Identification and Writer Retrieval Based on NetVLAD with Re- Ranking,” IET Biometrics, vol. 11, pp. 10–22, 2022

    2022

  17. [25]

    Writer Retrieval using Compact Convolutional Transformers and Net- MVLAD,

    M. Peer, F. Kleber, and R. Sablatnig, “Writer Retrieval using Compact Convolutional Transformers and Net- MVLAD,” in ICPR, 2022, pp. 1571–1578

    2022

  18. [26]

    Towards Writer Retrieval for Historical Datasets,

    ——, “Towards Writer Retrieval for Historical Datasets,” in ICDAR, 2023, pp. 411–427

    2023

  19. [27]

    Image Classification with the Fisher Vector: Theory and Practice,

    J. S ´anchez, F. Perronnin, T. Mensink, and J. Verbeek, “Image Classification with the Fisher Vector: Theory and Practice,” International Journal of Computer Vision, vol. 105, pp. 222–245, 2013

    2013

  20. [28]

    NetVLAD: CNN Architecture for Weakly Su- pervised Place Recognition,

    R. Arandjelovic, P. Gronat, A. Torii, T. Pajdla, and J. Sivic, “NetVLAD: CNN Architecture for Weakly Su- pervised Place Recognition,” in CVPR, 2016

    2016

  21. [29]

    Writer Identification and Re- trieval using a Convolutional Neural Network,

    S. Fiel and R. Sablatnig, “Writer Identification and Re- trieval using a Convolutional Neural Network,” in CAIP, 2015, pp. 26–37

    2015

  22. [30]

    Writer Identi- fication and Writer Retrieval using Vision Transformer for Forensic Documents,

    M. Koepf, F. Kleber, and R. Sablatnig, “Writer Identi- fication and Writer Retrieval using Vision Transformer for Forensic Documents,” in DAS, 2022, pp. 352–366

    2022

  23. [31]

    IAM-OnDB - An On-Line English Sentence Database Acquired from Handwritten Text on a Whiteboard,

    M. Liwicki and H. Bunke, “IAM-OnDB - An On-Line English Sentence Database Acquired from Handwritten Text on a Whiteboard,” in ICDAR, 2005, pp. 956–961 V ol. 2

    2005

  24. [32]

    IBM UB 1: A Dual Mode Unconstrained En- glish Handwriting Dataset,

    A. Shivram, C. Ramaiah, S. Setlur, and V . Govin- daraju, “IBM UB 1: A Dual Mode Unconstrained En- glish Handwriting Dataset,” in ICDAR, 2013, pp. 13–17

    2013

  25. [33]

    Automatic Writer Identification Framework for Online Handwritten Documents using Character Prototypes,

    G. X. Tan, C. Viard-Gaudin, and A. C. Kot, “Automatic Writer Identification Framework for Online Handwritten Documents using Character Prototypes,” Pattern Recog- nition, vol. 42, no. 12, pp. 3313–3323, 2009

    2009

  26. [34]

    Individuality of Alphabet Knowledge in Online Writer Identification,

    ——, “Individuality of Alphabet Knowledge in Online Writer Identification,” International Journal on Docu- ment Analysis and Recognition , vol. 13, pp. 147–157, 2010

    2010

  27. [35]

    Fusing Asynchronous Feature Streams for On-line Writer Identification,

    A. Schlapbach and H. Bunke, “Fusing Asynchronous Feature Streams for On-line Writer Identification,” in ICDAR, vol. 1, 2007, pp. 103–107

    2007

  28. [36]

    A Subtractive Clustering Scheme for Text-Independent Online Writer Identifica- tion,

    G. Singh and S. Sundaram, “A Subtractive Clustering Scheme for Text-Independent Online Writer Identifica- tion,” in ICDAR, 2015, pp. 311–315

    2015

  29. [37]

    Online Writer Identification using Sparse Coding and Histogram Based Descriptors,

    I. Dwivedi, S. Gupta, V . Venugopal, and S. Sundaram, “Online Writer Identification using Sparse Coding and Histogram Based Descriptors,” in ICFHR, 2016, pp. 572– 577

    2016

  30. [38]

    Dissimilarity Gaussian Mixture Models for Efficient Of- fline Handwritten Text-Independent Identification Using SIFT and RootSIFT Descriptors,

    F. A. Khan, F. Khelifi, M. A. Tahir, and A. Bouridane, “Dissimilarity Gaussian Mixture Models for Efficient Of- fline Handwritten Text-Independent Identification Using SIFT and RootSIFT Descriptors,” IEEE Transactions on Information Forensics and Security , vol. 14, no. 2, pp....

    2019

  31. [39]

    Encoding Pathlet and SIFT IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY , VOL. 19, 2024 14 Features with Bagged VLAD for Historical Writer Iden- tification,

    S. Lai, Y . Zhu, and L. Jin, “Encoding Pathlet and SIFT IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY , VOL. 19, 2024 14 Features with Bagged VLAD for Historical Writer Iden- tification,” IEEE Transactions on Information Forensics and Security, vol. 15, pp. 3553–3566, 2020

    2024

  32. [40]

    A Writer Identification System for On-Line Whiteboard Data,

    A. Schlapbach, M. Liwicki, and H. Bunke, “A Writer Identification System for On-Line Whiteboard Data,” Pattern Recognition, vol. 41, no. 7, pp. 2381–2397, 2008

    2008

  33. [41]

    Online Writer Identi- fication with Sparse Coding-Based Descriptors,

    V . Venugopal and S. Sundaram, “Online Writer Identi- fication with Sparse Coding-Based Descriptors,” IEEE Transactions on Information Forensics and Security , vol. 13, no. 10, pp. 2538–2552, 2018

    2018

  34. [42]

    DeepWriter: A Multi-Stream Deep CNN for Text-Independent Writer Identification,

    L. Xing and Y . Qiao, “DeepWriter: A Multi-Stream Deep CNN for Text-Independent Writer Identification,” in ICFHR, 2016, pp. 584–589

    2016

  35. [43]

    PS-LSTM: Capturing Essen- tial Sequential Online Information with Path Signature and LSTM for Writer Identification,

    M. Liu, L. Jin, and Z. Xie, “PS-LSTM: Capturing Essen- tial Sequential Online Information with Path Signature and LSTM for Writer Identification,” in ICDAR, vol. 01, 2017, pp. 664–669

    2017

  36. [44]

    End- to-End Online Writer Identification with Recurrent Neu- ral Network,

    X.-Y . Zhang, G.-S. Xie, C.-L. Liu, and Y . Bengio, “End- to-End Online Writer Identification with Recurrent Neu- ral Network,” IEEE Transactions on Human-Machine Systems, vol. 47, no. 2, pp. 285–292, 2017

    2017

  37. [45]

    Deep Adaptive Learning for Writer Identification Based on Single Handwritten Word Images,

    S. He and L. Schomaker, “Deep Adaptive Learning for Writer Identification Based on Single Handwritten Word Images,” Pattern Recognition, vol. 88, pp. 64–74, 2019

    2019

  38. [46]

    FragNet: Writer Identification using Deep Frag- ment Networks,

    ——, “FragNet: Writer Identification using Deep Frag- ment Networks,” IEEE Transactions on Information Forensics and Security, vol. 15, pp. 3013–3022, 2020

    2020

  39. [47]

    RSTC: A New Residual Swin Transformer for Offline Word-Level Writer Identification,

    P. Zhang, “RSTC: A New Residual Swin Transformer for Offline Word-Level Writer Identification,” IEEE Access, vol. 10, pp. 57 452–57 460, 2022

    2022

  40. [48]

    Attention is All You Need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. u. Kaiser, and I. Polosukhin, “Attention is All You Need,” in NeurIPS, vol. 30, 2017

    2017

  41. [49]

    An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale,

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weis- senborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Min- derer, G. Heigold, S. Gelly, J. Uszkoreit, and N. Houlsby, “An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale,” in ICLR, 2021

    2021

  42. [50]

    Swin Transformer: Hierarchical Vision Transformer using Shifted Windows,

    Z. Liu, Y . Lin, Y . Cao, H. Hu, Y . Wei, Z. Zhang, S. Lin, and B. Guo, “Swin Transformer: Hierarchical Vision Transformer using Shifted Windows,” in ICCV, 2021, pp. 10 012–10 022

    2021

  43. [51]

    Escaping the Big Data Paradigm with Compact Transformers,

    A. Hassani, S. Walton, N. Shah, A. Abuduweili, J. Li, and H. Shi, “Escaping the Big Data Paradigm with Compact Transformers,” arXiv preprint arXiv:2104.05704 , 2021

    2021 arXiv

  44. [52]

    Online Text-independent Writer Identi- fication Based on Temporal Sequence and Shape Codes,

    B. Li and T. Tan, “Online Text-independent Writer Identi- fication Based on Temporal Sequence and Shape Codes,” in ICDAR, 2009, pp. 931–935

    2009

  45. [53]

    Squeeze-and-Excitation Networks,

    J. Hu, L. Shen, and G. Sun, “Squeeze-and-Excitation Networks,” in CVPR, 2018

    2018

  46. [54]

    Rectified Linear Units Im- prove Restricted Boltzmann Machines,

    V . Nair and G. E. Hinton, “Rectified Linear Units Im- prove Restricted Boltzmann Machines,” in ICML, 2010, pp. 807–814

    2010

  47. [55]

    Self-Normalizing Neural Networks,

    G. Klambauer, T. Unterthiner, A. Mayr, and S. Hochre- iter, “Self-Normalizing Neural Networks,” in NeurIPS, vol. 30, 2017

    2017

  48. [56]

    Deep Residual Learning for Image Recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep Residual Learning for Image Recognition,” in CVPR, 2016

    2016

  49. [57]

    Circle Loss: A Unified Perspective of Pair Similarity Optimization,

    Y . Sun, C. Cheng, Y . Zhang, C. Zhang, L. Zheng, Z. Wang, and Y . Wei, “Circle Loss: A Unified Perspective of Pair Similarity Optimization,” in CVPR, 2020

    2020

  50. [58]

    Joint Detection and Identification Feature Learning for Person Search,

    T. Xiao, S. Li, B. Wang, L. Lin, and X. Wang, “Joint Detection and Identification Feature Learning for Person Search,” in CVPR, 2017

    2017

  51. [59]

    ABD-Net: Attentive but Diverse Person Re-Identification,

    T. Chen, S. Ding, J. Xie, Y . Yuan, W. Chen, Y . Yang, Z. Ren, and Z. Wang, “ABD-Net: Attentive but Diverse Person Re-Identification,” in ICCV, 2019

    2019

  52. [60]

    EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks,

    M. Tan and Q. Le, “EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks,” in ICML, vol. 97, 2019, pp. 6105–6114

    2019

  53. [61]

    A ConvNet for the 2020s,

    Z. Liu, H. Mao, C.-Y . Wu, C. Feichtenhofer, T. Darrell, and S. Xie, “A ConvNet for the 2020s,” in CVPR, 2022, pp. 11 976–11 986

    2022

  54. [62]

    HorNet: Efficient High-Order Spatial Interactions with Recursive Gated Convolutions,

    Y . Rao, W. Zhao, Y . Tang, J. Zhou, S. N. Lim, and J. Lu, “HorNet: Efficient High-Order Spatial Interactions with Recursive Gated Convolutions,” inNeurIPS, vol. 35, 2022, pp. 10 353–10 366

    2022

  55. [63]

    Run, Don’t Walk: Chasing Higher FLOPS for Faster Neural Networks,

    J. Chen, S.-h. Kao, H. He, W. Zhuo, S. Wen, C.-H. Lee, and S.-H. G. Chan, “Run, Don’t Walk: Chasing Higher FLOPS for Faster Neural Networks,” in CVPR, 2023, pp. 12 021–12 031

    2023

  56. [64]

    MSDS: A Large- Scale Chinese Signature and Token Digit String Dataset for Handwriting Verification,

    P. Zhang, J. Jiang, Y . Liu, and L. Jin, “MSDS: A Large- Scale Chinese Signature and Token Digit String Dataset for Handwriting Verification,” in NeurIPS, vol. 35, 2022, pp. 36 507–36 519

    2022

  57. [65]

    Decoupled Weight Decay Regularization,

    I. Loshchilov and F. Hutter, “Decoupled Weight Decay Regularization,” in ICLR, 2019

    2019

  58. [66]

    PyTorch: An Imperative Style, High-Performance Deep Learning Library,

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. Kopf, E. Yang, Z. DeVito, M. Rai- son, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, and S. Chintala, “PyTorch: An Imperative Style, High-...

    2019

  59. [67]

    Deep Neural Network for Online Writer Identification using Beta-Elliptic Model,

    T. Dhieb, W. Ouarda, H. Boubaker, and A. M. Alimi, “Deep Neural Network for Online Writer Identification using Beta-Elliptic Model,” in IJCNN, 2016, pp. 1863– 1870

    2016

  60. [68]

    ICFHR 2018 – Competition on Vietnamese Online Handwrit- ten Text Recognition using HANDS-VNOnDB (VO- HTR2018),

    H. T. Nguyen, C. T. Nguyen, and M. Nakagawa, “ICFHR 2018 – Competition on Vietnamese Online Handwrit- ten Text Recognition using HANDS-VNOnDB (VO- HTR2018),” in ICFHR, 2018, pp. 494–499

    2018

  61. [69]

    A New Database for Online Handwritten Mongolian Word Recognition,

    L.-L. Ma, J. Liu, and J. Wu, “A New Database for Online Handwritten Mongolian Word Recognition,” in ICPR, 2016, pp. 1131–1136

    2016

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.