Pith. sign in

REVIEW 3 major objections 6 minor 56 references

Sign-IDD: Iconicity Disentangled Diffusion for Sign Language Production

T0 review · 3 major / 6 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 Representing sign poses as bone directions and lengths, instead of raw joint coordinates, makes a diffusion model produce more accurate and semantically faithful sign poses.

desk verdict Solid, incremental sign-language pose generation with a real evaluation weakness: the PHOENIX14T ground-truth poses are pseudo-labels, and the paper's headline claims rest on them. read the letter →

arxiv 2412.13609 v2 pith:FAM4I7QL submitted 2024-12-18 cs.CV cs.MM

Shengeng Tang , Jiayi He , Dan Guo , Yanyan Wei , Feng Li , Richang Hong This is my paper · ORCID
classification cs.CVcs.MM
keywords signlanguageproductiongloss-to-posediffusionmodels3DposegenerationbonerepresentationdisentanglementskeletonconstraintsPHOENIX14T
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

Sign Language Production from glosses usually treats a pose as a set of independent 3D joint coordinates and fits those coordinates directly. This paper argues that such a representation ignores the relative spatial relationships between joints, which are what make signs readable. It proposes Sign-IDD, a diffusion framework that first disentangles each 3D pose into bones, each described by a 3D directional vector and a 1D length, and then uses those bone attributes to control pose generation alongside the original coordinates. The claim, supported by experiments on PHOENIX14T and USTC-CSL, is that adding this bone-level constraint yields poses that are semantically closer to the intended signs and show fewer finger overlaps and limb-orientation errors. If correct, this makes text-to-sign production more reliable without requiring new data.

What carries the argument

The workhorse is the 4D bone representation obtained by iconicity disentanglement. For each parent-child pair of joints along the body topology, the representation is $q^* = (q_c - q_p)/||q_c - q_p||_2$ for the 3D unit direction and $m = ||q_c - q_p||_2$ for the length, with the neck as root joint; this turns absolute joint coordinates into relative bone descriptors. The ACD module is the second piece: it concatenates the 3D and 4D pose streams, injects gloss embeddings through multi-head attention, then uses an attribute separation layer to split features back into coordinate and attribute streams and an attribute control layer to let bone attributes refine the coordinate features during denoising. A bone constraint loss supervises the predicted directions directly. Together these components give the model an explicit handle on bone orientation and length, which is what carries the argument.

What would settle it

Generate sign poses on a dataset with true 3D labels (or motion capture) using both Sign-IDD and a 3D-coordinate-only diffusion baseline and compare MPJPE/MPJAE; if the bone-based model does not still win on clean ground truth, the reported advantage is an artifact of pseudo-label noise. A simpler version: train with USTC-CSL real poses and check whether the improvement over baseline matches the PHOENIX14T improvement.

Watch

Extended reading notes

Core claim

The central claim is that converting a 3D joint pose into a 4D bone representation, where every bone carries the normalized direction from parent to child joint and the Euclidean distance between them, exposes the skeletal structure that coordinate regression hides. On PHOENIX14T TEST, Sign-IDD reports BLEU-1 24.80, WER 76.66, FID 2.46, MPJPE 47.19, and on USTC-CSL it reports BLEU-1 65.26 and WER 50.15, beating the compared baselines on most metrics. The paper attributes these gains to two components: the Iconicity Disentanglement (ID) module, which derives the 4D bone representation and feeds it together with the 3D coordinates into the denoiser, and the Attribute Controllable Diffusion (ACD) module, which separates bone direction and length features from coordinate features and uses them as control signals during reverse diffusion. It also adds a bone-orientation loss $L_{bone}$ to the usual joint loss. The paper's conclusion is that modeling both joint positions and their relative associations is what improves pose accuracy and semantic consistency.

Load-bearing premise

All PHOENIX14T results rest on computer-estimated 3D poses standing in for true sign poses; if those pseudo-labels are systematically biased, the reported pose-accuracy improvements are not measurements of real pose accuracy.

Editorial extensions

If this is right

  • Sign pose generators can be made more precise by explicitly modeling bone orientation and length, not only joint coordinates; the reported gains on BLEU, WER, FID, MPJPE, and MPJAE point to fewer semantic errors and fewer distorted hands and limbs.
  • The 4D bone representation is a cheap side channel that works with existing gloss-to-pose diffusion architectures, so other SLP systems could adopt it without changing their overall training regime.
  • The ablation results show that both the ID and ACD modules contribute beyond the diffusion baseline, and that the bone constraint weight $\lambda = 0.1$ matters: too small or too large a weight degrades performance, so there is a practical tuning range for skeletal supervision.
  • The method is tested on a German sign corpus and a Chinese sign corpus, suggesting the bone-disentanglement benefit is not language-specific.

Reading between the lines

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

  • Because PHOENIX14T has no real pose labels, the paper's MPJPE/MPJAE numbers on that corpus measure error against pseudo-ground-truth produced by OpenPose and a lifting model; a fair test of absolute pose accuracy would require clean 3D labels, for example motion capture or manual annotation.
  • The same direction-plus-length disentanglement could transfer to other articulated motion generation tasks, such as hand gesture synthesis or human motion prediction, where relative joint geometry carries much of the meaning.
  • A control experiment that gives a 3D-coordinate-only baseline the same extra parameters would show whether the gains come from the disentangled representation itself or merely from added network capacity; the paper does not isolate this.
  • The USTC-CSL results, which use real pose labels, are the cleaner evidence for the method's pose-level claims; if future work focuses there, the bone-representation advantage can be verified without pseudo-label confounds.
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

3 major / 6 minor

Summary. The paper proposes Sign-IDD, a gloss-to-pose diffusion model for sign language production. Instead of regressing only 3D joint coordinates, it converts poses into a '4D bone representation' of unit direction vectors and bone lengths (Eq. 4), concatenates this 4D representation with the 3D pose, and feeds the result through an attribute separation/control module inside a diffusion denoiser. Training uses a joint loss (Eq. 15) and a bone-orientation loss (Eq. 16). The method is evaluated on PHOENIX14T and USTC-CSL with BLEU/ROUGE/WER via NSLT back-translation plus FID, MPJPE, and MPJAE, and the paper reports improvements over several published baselines and over its own ablations.

Significance. If the reported results are reliable, the paper makes a modest but useful contribution: it shows that an explicit relative/bone representation can improve pose generation in SLP, and it provides a clean, parameter-light modification of a diffusion backbone. The code is released, the paper includes a ground-truth row to calibrate the back-translation metrics, and the ablations separately test both the ID and ACD components. The main uncertainties are the quality of the pseudo-label ground truth on PHOENIX14T and the absence of variance estimates; these currently prevent me from judging whether the claimed gains are due to the proposed mechanism or to fitting label artifacts.

major comments (3)
  1. [Implementation Details; Eqs. (15)-(16); Table 1] The PHOENIX14T pose ground truth is produced by OpenPose 2D keypoints lifted with the Zelinka-Kanis skeletal correction model, because the dataset has no native pose labels. These pseudo-labels serve simultaneously as the training targets for Ljoint and Lbone and as the reference for MPJPE, MPJAE, FID, and the NSLT-based BLEU/WER scores. Any systematic bias in the pseudo-labels, especially in hand articulation and limb orientation, is therefore baked into both the objective and the measuring stick. The paper itself notes that the ground truth contains 'inaccurate labels' and 'motion blur' in Figure 4. To support the central attribution claim, please add ablations and comparisons on a dataset with native pose annotations (or otherwise verify against independent annotations), report a quantitative sensitivity analysis using an independent pose estimator on a subset of PHOENIX14T, or provide a human evaluation of pose correctness.
  2. [Tables 3 and 4] All ablations are single-run and are reported only on PHOENIX14T. The increments attributed to the modules are small in absolute terms (e.g., Base+ID vs Base+ID+ACD on TEST BLEU-1: 23.46 vs 24.80, and WER 77.53 vs 76.66), so without multiple seeds, error bars, or significance tests the claim that ID and ACD are each responsible for the gains is not statistically supported. Please report run-to-run variance or significance tests, and, if possible, repeat the ablations on USTC-CSL.
  3. [Table 1] The comparison is incomplete in a way that affects the headline claim. NAT-ATM, NAT-EAM, and G2P-DDM have no FID/MPJPE/MPJAE entries, DET* has no WER/FID/MPJPE/MPJAE, and GCDM lacks FID/MPJPE/MPJAE on both DEV and TEST. The abstract and conclusions claim that the method 'outperforms SOTA methods,' but this is only partially supported by the metrics that are actually reported. Please either obtain and report these metrics with a shared protocol (including the same NSLT model and pose representation) or explicitly restrict the claim to the metrics where all methods are compared.
minor comments (6)
  1. [Eq. (6) and Reverse Process] The fused representation in Eq. (6) is written as [pt, p′t] ∈ R7, but pt has J joints while p′t has B = J − 1 bones according to Eq. (4). Please state explicitly that a zero 4D root bone is included so that the two sequences have the same length; otherwise the concatenation is undefined.
  2. [Eq. (12) and inference procedure] The description of the inference timestep is confusing: 'Initiated at T, the timestep for each iteration is computed as T = T − (1−i/I)' reuses the symbol T. Please clarify the schedule and the role of i.
  3. [Introduction and Related Work] The term 'iconicity' is used in a nonstandard way. In sign linguistics, iconicity refers to a resemblance between form and meaning, not to invariance of semantics under a coordinate transformation. Please either define the term clearly in the paper or use a more neutral name such as 'bone disentanglement.'
  4. [Table 4] There is a typo in the parameter row of Table 4: 'angi = 5' should be 'i = 5'.
  5. [Figure 1] Figure 1 labels both the traditional 3D representation and the proposed 4D representation as '(b)', which is confusing; the second panel should be '(c)'.
  6. [USTC-CSL comparison] The paper states that several typical solutions were tested 'under a fair setting' on USTC-CSL, but does not describe this setting. Please provide implementation details for the reimplemented baselines (e.g., training protocol, pose representation, and any hyperparameter tuning).

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: improvements are measured against public benchmarks and published baselines; only minor reuse of the authors' GCDM gloss encoder.

full rationale

Sign-IDD's central claim is that adding bone direction/length supervision (ID and ACD) improves sign pose generation. This is evaluated on PHOENIX14T and USTC-CSL against published methods; the reported gains are not derived from the method's own parameters. The 4D representation is the deterministic map q*=(qc-qp)/||qc-qp||2, m=||qc-qp||2 (Eq. 4), so it contains no new information beyond the 3D input; this is a reparameterized training target, not a fitted parameter renamed as a prediction, and it does not make the evaluation metric equal to the training loss by construction. The Gloss Encoder is cited to the authors' own GCDM (Tang et al. 2024), but it is a reusable component used in both the baseline and the full model, so the improvement of Base+ID+ACD over Base cannot reduce to that citation. The PHOENIX14T pseudo-label issue (OpenPose + Zelinka/Kanis lifting) is a data-quality limitation that affects external validity, but since the same pseudo-labels are used consistently for training and for all compared methods, it is not circular in the derivation sense. No uniqueness theorem or ansatz is imported from the authors' prior work to forbid alternatives. Therefore no equation in the paper exhibits the claimed result as equivalent to its own input; circularity score is low.

Assumptions & free parameters 3 free parameters · 3 assumptions · 1 invented entities

The central claim relies on pseudo-ground-truth 3D poses, back-translation as an evaluation proxy, and gloss as semantic input, plus three tuned hyperparameters. No new physical entity is introduced; the 4D representation is a deterministic transform of the 3D joints.

free parameters (3)
  • lambda (bone loss weight) = 0.1
    Selected by ablation on PHOENIX14T (Table 4); the headline numbers use this value.
  • t (diffusion timesteps) = 1000
    Selected by ablation on PHOENIX14T (Table 4).
  • i (inference sampling steps) = 5
    Selected by ablation on PHOENIX14T (Table 4).
assumptions (3)
  • domain assumption PHOENIX14T has no pose labels; OpenPose 2D detections lifted to 3D by Zelinka and Kanis (2020) are treated as ground-truth 3D poses for training and evaluation.
    Invoked in 'Implementation Details'; all PHOENIX14T numbers depend on this estimate.
  • domain assumption Back-translation scores from the NSLT model are a valid measure of semantic quality of generated sign poses.
    Metrics BLEU, ROUGE, and WER are computed through an SLT model trained on the same data, so errors in the back-translator are confounded with pose quality.
  • domain assumption Gloss sequences are an adequate semantic conditioning signal for producing signs.
    The whole G2P pipeline assumes gloss input captures the meaning to be signed; this is the task definition used by prior work.
invented entities (1)
  • 4D bone representation
    purpose: Auxiliary pose representation that encodes each bone as a 3D direction vector plus a 1D length, used as extra supervision and as input to the attribute control layer.
    It is deterministically derived from the 3D joint coordinates by Eq. 4, so it carries no information beyond the original pose and no external measurement supports it.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Sign-IDD: Iconicity Disentangled Diffusion for Sign Language Production." pith.science (2026). https://pith.science/paper/FAM4I7QL

@misc{pith2026241213609,
  author       = {Pith},
  title        = {Pith review of: Sign-IDD: Iconicity Disentangled Diffusion for Sign Language Production},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FAM4I7QL}},
  note         = {Machine review of arXiv:2412.13609}
}
read the original abstract

Sign Language Production (SLP) aims to generate semantically consistent sign videos from textual statements, where the conversion from textual glosses to sign poses (G2P) is a crucial step. Existing G2P methods typically treat sign poses as discrete three-dimensional coordinates and directly fit them, which overlooks the relative positional relationships among joints. To this end, we provide a new perspective, constraining joint associations and gesture details by modeling the limb bones to improve the accuracy and naturalness of the generated poses. In this work, we propose a pioneering iconicity disentangled diffusion framework, termed Sign-IDD, specifically designed for SLP. Sign-IDD incorporates a novel Iconicity Disentanglement (ID) module to bridge the gap between relative positions among joints. The ID module disentangles the conventional 3D joint representation into a 4D bone representation, comprising the 3D spatial direction vector and 1D spatial distance vector between adjacent joints. Additionally, an Attribute Controllable Diffusion (ACD) module is introduced to further constrain joint associations, in which the attribute separation layer aims to separate the bone direction and length attributes, and the attribute control layer is designed to guide the pose generation by leveraging the above attributes. The ACD module utilizes the gloss embeddings as semantic conditions and finally generates sign poses from noise embeddings. Extensive experiments on PHOENIX14T and USTC-CSL datasets validate the effectiveness of our method. The code is available at: https://github.com/NaVi-start/Sign-IDD.

Figures

Figures reproduced from arXiv: 2412.13609 by the authors.

Figure 1
Figure 1. (a) Problems caused by using only traditional 3D [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of our framework. Given a gloss sequence, Sign-IDD generates a coherent sign pose video guided by gloss [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The main components of our ACD module. Forward Process As illustrated in [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Visualization examples of generated poses on PHOENIX14T (top) and USTC-CSL (bottom). We compare Sign-IDD [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Visualization results of Sign-IDD and Base, which [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

56 extracted references · 42 canonical work pages

  1. [1]

    A.; Ververas, E.; Sun, G.; Deng, J.; and Zafeiriou, S

    Baltatzis, V.; Potamias, R. A.; Ververas, E.; Sun, G.; Deng, J.; and Zafeiriou, S. 2024. Neural Sign Actors: A Diffusion Model for 3D Sign Language Production from Text. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, 1985--1995

    work page 2024

  2. [2]

    Cai, Q.; Hu, X.; Hou, S.; Yao, L.; and Huang, Y. 2024. Disentangled Diffusion-Based 3D Human Pose Estimation with Hierarchical Spatial and Temporal Denoiser. In AAAI Conference on Artificial Intelligence, 882--890

    work page 2024

  3. [3]

    C.; Hadfield, S.; Koller, O.; Ney, H.; and Bowden, R

    Camgoz, N. C.; Hadfield, S.; Koller, O.; Ney, H.; and Bowden, R. 2018. Neural Sign Language Translation. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, 7784--7793

    work page 2018

  4. [4]

    C.; Koller, O.; Hadfield, S.; and Bowden, R

    Camgoz, N. C.; Koller, O.; Hadfield, S.; and Bowden, R. 2020. Sign Language Transformers: Joint End-to-End Sign Language Recognition and Translation. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, 10023--10033

    work page 2020

  5. [5]

    Cao, Z.; Simon, T.; Wei, S.-E.; and Sheikh, Y. 2017. Realtime Multi-Person 2d Pose Estimation Using Part Affinity Fields. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, 7291--7299

    work page 2017

  6. [6]

    Cui, R.; Liu, H.; and Zhang, C. 2017. Recurrent Convolutional Neural Networks for Continuous Sign Language Recognition by Staged Optimization. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, 7361--7369

    work page 2017

  7. [7]

    Fang, S.; Sui, C.; Zhang, X.; and Tian, Y. 2023. SignDiff: Learning Diffusion Models for American Sign Language Production. arXiv Preprint ArXiv:2308.16082, arXiv--2308

    arXiv 2023

  8. [8]

    Goodfellow, I.; Pouget-Abadie, J.; Mirza, M.; Xu, B.; Warde-Farley, D.; Ozair, S.; Courville, A.; and Bengio, Y. 2014. Generative Adversarial Nets. Advances in Neural Information Processing Systems, 2672–2680

    work page 2014

Show all 56 references
  1. [9]

    Guo, D.; Tang, S.; Hong, R.; and Wang, M. 2021. Sign Language Recognition. Multimedia for Accessible Human Computer Interfaces, 23--59

    2021

  2. [10]

    Guo, D.; Tang, S.; and Wang, M. 2019. Connectionist Temporal Modeling of Video and Language: A Joint Model for Translation and Sign Labeling. In International Joint Conference on Artificial Intelligence, 751--757

    2019

  3. [11]

    Guo, D.; Zhou, W.; Li, H.; and Wang, M. 2017. Online Early-Late Fusion Based on Adaptive HMM for Sign Language Recognition. ACM Transactions on Multimedia Computing, Communications, and Applications, 1--18

    2017

  4. [12]

    Guo, D.; Zhou, W.; Li, H.; and Wang, M. 2018. Hierarchical LSTM for Sign Language Translation. In AAAI Conference on Artificial Intelligence, 6845--6852

    2018

  5. [13]

    Guo, M.; He, J.; Tang, S.; Wang, Z.; and Cheng, L. 2024. Shaping a Stabilized Video by Mitigating Unintended Changes for Concept-Augmented Video Editing. arXiv preprint arXiv:2410.12526

    2024 arXiv

  6. [14]

    Ho, J.; Jain, A.; and Abbeel, P. 2020. Denoising Diffusion Probabilistic Models. Advances in Neural Information Processing Systems, 6840--6851

    2020

  7. [15]

    Ho, J.; Salimans, T.; Gritsenko, A.; Chan, W.; Norouzi, M.; and Fleet, D. J. 2022. Video Diffusion Models. Advances in Neural Information Processing Systems, 8633--8646

    2022

  8. [16]

    Huang, J.; Zhou, W.; Zhang, Q.; Li, H.; and Li, W. 2018. Video-Based Sign Language Recognition without Temporal Segmentation. In AAAI Conference on Artificial Intelligence, 2257--2264

    2018

  9. [17]

    Huang, R.; Lam, M.; Wang, J.; Su, D.; Yu, D.; Ren, Y.; and Zhao, Z. 2022. FastDiff: A Fast Conditional Diffusion Model for High-Quality Speech Synthesis. In International Joint Conference on Artificial Intelligence, 4157--4163

    2022

  10. [18]

    Huang, W.; Pan, W.; Zhao, Z.; and Tian, Q. 2021. Towards Fast and High-Quality Sign Language Production. In ACM International Conference on Multimedia, 3172--3181

    2021

  11. [19]

    P.; and Ba, J

    Kingma, D. P.; and Ba, J. 2015. Adam: A Method for Stochastic Optimization. In International Conference on Learning Representations, 1--15

    2015

  12. [20]

    P.; and Welling, M

    Kingma, D. P.; and Welling, M. 2014. Auto-Encoding Variational Bayes. Stat, 1

    2014

  13. [21]

    Koller, O. 2020. Quantitative Survey of the State of the Art in Sign Language Recognition. arXiv Preprint ArXiv:2008.09918, arXiv--2008

    2020 arXiv

  14. [22]

    Li, F.; Cong, R.; Wu, J.; Bai, H.; Wang, M.; and Zhao, Y. 2024 a . SRConvNet: A Transformer-Style ConvNet for Lightweight Image Super-Resolution. International Journal of Computer Vision, 1--17

    2024

  15. [23]

    Li, F.; Wu, Y.; Li, A.; Bai, H.; Cong, R.; and Zhao, Y. 2024 b . Enhanced Video Super-Resolution Network towards Compressed Data. ACM Transactions on Multimedia Computing, Communications and Applications, 1--21

    2024

  16. [24]

    S.; and Hashimoto, T

    Li, X.; Thickstun, J.; Gulrajani, I.; Liang, P. S.; and Hashimoto, T. B. 2022. Diffusion-LM Improves Controllable Text Generation. Advances in Neural Information Processing Systems, 4328--4343

    2022

  17. [25]

    Makhzani, A.; Shlens, J.; Jaitly, N.; Goodfellow, I.; and Frey, B. 2015. Adversarial Autoencoders. arXiv e-prints, arXiv--1511

    2015

  18. [26]

    P.; and Jawahar, C

    Mazumder, S.; Mukhopadhyay, R.; Namboodiri, V. P.; and Jawahar, C. 2021. Translating Sign Language Videos to Talking Faces. In Indian Conference on Computer Vision, Graphics and Image Processing, 1--10

    2021

  19. [27]

    G.; Stumbo, M.; Berke, L.; Bialek, M.; and Thomas, F

    McDonald, J.; Wolfe, R.; Schnepp, J.; Hochgesang, J.; Jamrozik, D. G.; Stumbo, M.; Berke, L.; Bialek, M.; and Thomas, F. 2016. An Automated Technique for Real-Time Production of Lifelike Animations of American Sign Language. Universal Access in the Information Society, 551--566

    2016

  20. [28]

    Moryossef, A.; Yin, K.; Neubig, G.; and Goldberg, Y. 2021. Data Augmentation for Sign Language Gloss Translation. In International Workshop on Automatic Translation for Signed and Spoken Languages, 1--11

    2021

  21. [29]

    K.; and Dingemanse, M

    Nielsen, A. K.; and Dingemanse, M. 2021. Iconicity in Word Learning and Beyond: A Critical Review. Language and Speech, 52--72

    2021

  22. [30]

    Orbay, A.; and Akarun, L. 2020. Neural Sign Language Translation by Learning Tokenization. In IEEE International Conference on Automatic Face and Gesture Recognition, 222--228

    2020

  23. [31]

    Othman, A.; and Jemni, M. 2011. Statistical Sign Language Machine Translation: from English Written Text to American Sign Language Gloss. International Journal of Computer Science Issues, 65

    2011

  24. [32]

    C.; and Bowden, R

    Saunders, B.; Camgoz, N. C.; and Bowden, R. 2020 a . Adversarial Training for Multi-Channel Sign Language Production. In British Machine Vision Conference, 1--15

    2020

  25. [33]

    C.; and Bowden, R

    Saunders, B.; Camgoz, N. C.; and Bowden, R. 2020 b . Progressive Transformers for End-to-End Sign Language Production. In European Conference on Computer Vision, 687--705

    2020

  26. [34]

    C.; and Bowden, R

    Saunders, B.; Camgoz, N. C.; and Bowden, R. 2021 a . Continuous 3d Multi-Channel Sign Language Production via Progressive Transformers and Mixture Density Networks. International journal of computer vision, 2113--2135

    2021

  27. [35]

    C.; and Bowden, R

    Saunders, B.; Camgoz, N. C.; and Bowden, R. 2021 b . Mixed Signals: Sign Language Production via A Mixture of Motion Primitives. In IEEE/CVF International Conference on Computer Vision, 1919--1929

    2021

  28. [36]

    C.; and Bowden, R

    Saunders, B.; Camgoz, N. C.; and Bowden, R. 2022 a . Signing at Scale: Learning to Co-Articulate Signs for Large-Scale Photo-Realistic Sign Language Production. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, 5141--5151

    2022

  29. [37]

    C.; and Bowden, R

    Saunders, B.; Camgoz, N. C.; and Bowden, R. 2022 b . Skeletal Graph Self-Attention: Embedding a Skeleton Inductive Bias into Sign Language Production. In International Workshop on Sign Language Translation and Avatar Technology, 95--102

    2022

  30. [38]

    Segouat, J. 2009. A Study of Sign Language Coarticulation. ACM Sigaccess Accessibility and Computing, 31--38

    2009

  31. [39]

    Shan, W.; Liu, Z.; Zhang, X.; Wang, Z.; Han, K.; Wang, S.; Ma, S.; and Gao, W. 2023. Diffusion-Based 3D Human Pose Estimation with Multi-Hypothesis Aggregation. In IEEE/CVF International Conference on Computer Vision, 14761--14771

    2023

  32. [40]

    Shang, S.; Shan, Z.; Liu, G.; Wang, L.; Wang, X.; Zhang, Z.; and Zhang, J. 2024. Resdiff: Combining Cnn and Diffusion Model for Image Super-Resolution. In AAAI Conference on Artificial Intelligence, 8975--8983

    2024

  33. [41]

    Sohl-Dickstein, J.; Weiss, E.; Maheswaranathan, N.; and Ganguli, S. 2015. Deep Unsupervised Learning Using Nonequilibrium Thermodynamics. In International Conference on Machine Learning, 2256--2265

    2015

  34. [42]

    Song, P.; Guo, D.; Yang, X.; Tang, S.; and Wang, M. 2024. Emotional Video Captioning With Vision-Based Emotion Interpretation Network. IEEE Transactions on Image Processing, 1122--1135

    2024

  35. [43]

    Song, P.; Guo, D.; Yang, X.; Tang, S.; Yang, E.; and Wang, M. 2023. Emotion-Prior Awareness Network for Emotional Video Captioning. In ACM International Conference on Multimedia, 589--600

    2023

  36. [44]

    C.; Hadfield, S.; and Bowden, R

    Stoll, S.; Camgoz, N. C.; Hadfield, S.; and Bowden, R. 2020. Text2Sign: Towards Sign Language Production Using Neural Machine Translation and Generative Adversarial Networks. International Journal of Computer Vision, 891--908

    2020

  37. [45]

    Tang, S.; Guo, D.; Hong, R.; and Wang, M. 2022 a . Graph-Based Multimodal Sequential Embedding for Sign Language Translation. IEEE Transactions on Multimedia, 4433--4445

    2022

  38. [46]

    Tang, S.; Hong, R.; Guo, D.; and Wang, M. 2022 b . Gloss Semantic-Enhanced Network with Online Back-Translation for Sign Language Production. In ACM International Conference on Multimedia, 5630--5638

    2022

  39. [47]

    Tang, S.; Xue, F.; Wu, J.; Wang, S.; and Hong, R. 2024. Gloss-Driven Conditional Diffusion Models for Sign Language Production. ACM Transactions on Multimedia Computing, Communications, and Applications

    2024

  40. [48]

    N.; Kaiser, L.; and Polosukhin, I

    Vaswani, A.; Shazeer, N.; Parmar, N.; Uszkoreit, J.; Jones, L.; Gomez, A. N.; Kaiser, L.; and Polosukhin, I. 2017. Attention is All You Need. In Advances in Neural Information Processing Systems, 6000--6010

    2017

  41. [49]

    Viegas, C.; Inan, M.; Quandt, L.; and Alikhani, M. 2023. Including Facial Expressions in Contextual Embeddings for Sign Language Generation. In Joint Conference on Lexical and Computational Semantics, 1--10

    2023

  42. [50]

    Wang, H.; Chai, X.; and Chen, X. 2019. A Novel Sign Language Recognition Framework Using Hierarchical Grassmann Covariance Matrix. IEEE Transactions on Multimedia, 2806--2814

    2019

  43. [51]

    Wei, Y.; Zhang, Y.; Li, K.; Wang, F.; Tang, S.; and Zhang, Z. 2024. Leveraging Vision-Language Prompts for Real-World Image Restoration and Enhancement. Computer Vision and Image Understanding, 104222

    2024

  44. [52]

    Wu, J.; Hong, R.; and Tang, S. 2024. Intermediary-Generated Bridge Network for RGB-D Cross-Modal Re-Identification. ACM Transactions on Intelligent Systems and Technology, 1--25

    2024

  45. [53]

    Xie, P.; Zhang, Q.; Taiying, P.; Tang, H.; Du, Y.; and Li, Z. 2024. G2P-DDM: Generating Sign Pose Sequence from Gloss Sequence with Discrete Diffusion Model. In AAAI Conference on Artificial Intelligence, 6234--6242

    2024

  46. [54]

    Zelinka, J.; and Kanis, J. 2020. Neural Sign Language Synthesis: Words Are Our Glosses. In IEEE/CVF Winter Conference on Applications of Computer Vision, 3395--3403

    2020

  47. [55]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all...

  48. [56]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

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