The GENEA Challenge 2026: A Large-Scale Disentangled Evaluation of Speech-Driven Gesture Generation on the Seamless Interaction Dataset

Gustav Eje Henter; Hendric Voss; Mihail Tsakov; Rajmund Nagy; Silvia Arellano Garc\'ia; Taras Kucherenko; Youngwoo Yoon

REVIEW 3 major objections 5 minor 11 references

The GENEA Challenge 2026: A Large-Scale Disentangled Evaluation of Speech-Driven Gesture Generation on the Seamless Interaction Dataset

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Human motion capture beats all five gesture-generation systems on every evaluated dimension.

desk verdict The new semantic mismatching evaluation is promising but likely confounded by visible articulation and timing cues; the T1–T3 results are solid. read the letter →

arxiv 2608.10839 v1 pith:J34DT35Y submitted 2026-08-11 cs.CV cs.GRcs.HCcs.SD

Rajmund Nagy , Silvia Arellano García , Hendric Voss , Mihail Tsakov , Taras Kucherenko , Youngwoo Yoon , Gustav Eje Henter This is my paper · ORCID

classification cs.CVcs.GRcs.HCcs.SD

keywords gesturegenerationspeech-drivenanimationdyadicinteractionsemanticgesturesmismatchingevaluationmotioncapturebenchmarkuserstudySeamlessdataset

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 reports the results of the fourth GENEA Challenge, a large-scale crowdsourced evaluation of five speech-driven gesture-generation systems trained on the Seamless Interaction dataset of dyadic conversations. Its central claim is that recorded human motion is substantially better than every submitted system on all four tested dimensions: motion realism, alignment with the input speech, responsiveness to the interlocutor, and semantic expressiveness of gestures. The mismatch-based evaluation methodology is designed so that an input-independent system should score zero, giving the results an interpretable floor; under that metric, all five systems score near zero on dyadic responsiveness and semantic expressiveness while motion capture reaches 65 and 77 percent respectively. The paper additionally proposes a new semantic mismatching methodology using the Grounded Gestures subset of the dataset, and argues that the dataset and procedure are suitable for benchmarking these harder capabilities in the future.

What carries the argument

The central mechanism is a set of mismatching evaluations in which the attribute under test is isolated by holding everything else fixed. T2 pairs the same motion with matched versus mismatched audio chosen to have similar voice-activity structure; T3 pairs matched and mismatched interlocutor contexts while keeping both speakers' audio and the interlocutor's motion identical; T4, the paper's new contribution, shows a single muted video of a monadic gesture and asks test-takers to pick which of two sentences the gestures express. These designs convert the evaluation into a discrimination task with an interpretable chance level of zero for an input-independent system, and pairwise Likert or forced-choice voting is aggregated into Elo ratings and appropriateness scores in the range [-1, 1].

What would settle it

A control study in which the 'mismatched' audio is matched to the original on prosody, voice, and recording conditions (not just voice-activity structure) should reproduce the 62 percent ceiling for motion capture; if test-takers cannot tell matched from such a near-twin audio, the T2 score is inflated by non-semantic acoustics. Similarly for T4, replacing the mismatched sentence with one of matched length and similar keywords should drop the 79 percent mocap identification if the test measures meaning rather than superficial cue matching.

Watch

Extended reading notes

Core claim

The paper's discovery is that, on the Seamless Interaction dataset, the gap between human motion and state-of-the-art learned gesture generation remains wide on every axis, and is largest on exactly the capabilities that motivate dyadic data. In pairwise realism comparisons the motion-capture reference wins 68 to 95 percent of matches; in speech-mismatching tests it posts a 62 percent appropriateness score against 32 percent for the top submission, with the rest near the zero floor; in dyadic mismatching it scores 65 percent with all submissions at chance; and in the new semantic mismatching study test-takers identify the correct sentence from mocap video 79 percent of the time, while the best system scores 8 percent. The paper treats these numbers as evidence that current systems cannot yet produce interlocutor-responsive or semantically expressive motion, and that the newly introduced text-mismatching procedure gives a valid, interpretable measure of that shortfall.

Load-bearing premise

The mismatched stimuli differ from the matched ones only in the attribute being measured, so that a preference for the matched video reflects dyadic responsiveness, speech alignment, or semantic expressiveness rather than some side difference such as voice, prosody, rendering, or motion quality.

Editorial extensions

If this is right

If the results hold, Seamless Interaction with the mismatching methodology becomes a benchmark on which progress in dyadic and semantic gesture generation can be measured against interpretable zero-floor and mocap-ceiling anchors.
Any future system must clear the demonstrated gap between 8 percent and 77 percent on semantic appropriateness before it can credibly claim meaning-aware gesture generation.
The near-chance dyadic scores indicate that models trained on dyadic data are not yet using the interlocutor's audio or motion at generation time in a way test-takers perceive as responsive.
The speech-alignment results put the best submission at roughly half the mocap ceiling, implying that state-of-the-art co-speech rhythm still lags human timing.

Reading between the lines

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

The reported near-chance scores for systems also depend on how representative the five submissions are of the field; the stronger system that previously reached mocap-level alignment on the leaderboard was not among the submissions, so the challenge may underestimate the current state of the art.
The semantic mismatching method's ceiling of 79 percent on mocap, not 100, suggests the grounded-gesture clips still contain a sizable fraction of gestures that test-takers cannot map to a specific sentence; the method measures recognizable expressivity, not all communicative content.
A stronger control would withhold all transcript information and instead compare matched versus mismatched texts with identical keywords or identical prosody, which would tell whether the 79 percent reflects lexical iconicity or speaker-specific delivery.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Colleague, the key thing to know: GENEA 2026 is the first large-scale attempt to evaluate semantic expressivity in gesture generation via text mismatching, and on that new axis the gap between motion capture and submissions is enormous (77% vs 8%). The evaluation is careful and mostly convincing for realism, speech alignment, and dyadic alignment. But the semantic result rests on an untested assumption, and I would not take the 79% vs 8% as strong evidence about semantics until the confound is addressed.

What is genuinely new: a fourth GENEA-style evaluation task using the Grounded Gestures subset, text mismatching rather than audio mismatching, applied to the Seamless Interaction dataset. The scale (23k votes, 869 test-takers) and the bootstrapped confidence intervals give the headline numbers real weight. The updated appropriateness scaling is an explicit linear transform of the 2023 formula, so there is no circularity. The authors also openly discuss system-level and study-level randomness, which is more than most challenge reports do.

The soft spot is T4. Test-takers see a muted video of the speaker and choose which of two sentences the gestures express. The matched sentence is the one actually spoken, with a marked target word. Visible mouth/jaw articulation and the timing/emphasis of that target word will both correlate with the matched condition for the mocap videos, and both are essentially absent from generated body-only motion. The mismatched sentence is drawn at random, so it will usually differ in length, stress, and target-word position. That means a test-taker could identify the matched transcript from low-level temporal or articulatory cues without reading any semantic content in the gestures. The paper reports no control that removes the face or masks articulation. Given that this is the paper's new methodology and the basis for a headline conclusion, the 79% vs 8% gap is exactly what such a confound would produce. The dyadic study has a similar but smaller issue: the 30-50% VAD overlap thresholds are empirically chosen, and the segments are manually curated, so the ceiling comparison is shaped by the selection.

The T1–T3 results stand mostly on their own, though the promises that votes/outputs and system descriptions will be released later mean the reproducibility is currently prospective. If the promised data ship, this will be a useful benchmark paper for gesture-generation evaluators. I would send it to review, and I would cite it for the T1–T3 findings; I would be cautious about citing the semantic result without the missing control.

Recommendation: accept the challenge report with revision, and require the authors to either report a face-masking/articulation-removal control for T4 or soften the semantic claim.

Referee Report

3 major / 5 minor

Summary. This paper reports the GENEA Challenge 2026, a large-scale crowdsourced evaluation of five speech-driven gesture-generation systems trained on the Seamless Interaction dataset. The authors describe four disentangled user studies: motion realism (E1), speech-motion alignment via audio mismatching (E2), dyadic alignment via interlocutor mismatching (E3), and a newly proposed semantic alignment task based on text mismatching (E4). Across more than 23,000 votes, the filtered motion-capture reference outperformed all submissions on every axis, with submissions near chance on speech, dyadic, and semantic alignment. The paper also introduces an updated appropriateness scaling centered at zero and releases the collected votes and outputs.

Significance. If the results hold, the paper makes a useful contribution as the first large-scale evaluation of gesture generation on the Seamless Interaction dataset and as a proposal for a semantic text-mismatching evaluation. Its strengths are the large vote counts, bootstrapped confidence intervals, pairwise disentangled design, JUICE justifications, and the public release of votes and outputs. However, the new semantic task is the most novel and headline-producing component, and its validity currently rests on an untested assumption that no non-semantic cues are available to test-takers; the semantic score also contains an apparent arithmetic inconsistency. The paper is therefore promising but needs additional control analyses before the central claims can be accepted.

major comments (3)

[§3.1.1 (T4), Figs. 13–14] The paper's headline semantic result—79% matched-text identification for motion capture versus no more than 8% for the submissions—is interpreted as semantic expressiveness, but the text-mismatching task as described does not isolate semantic content. In a single muted monadic video, two non-semantic cues are available if the renderings show the speaker's face, which the paper does not state: visible mouth/jaw articulation, and the timing/emphasis of the marked target word. Since the mismatched sentence is drawn at random, it will generally differ in length, prosody, and target-word position, so a test-taker can identify the matched sentence by low-level temporal or articulatory matching even if the gestures carry no meaning. These cues are perfectly correlated with the matched condition for the motion-capture reference and are not produced by body-only gesture generators, which is precisely the pattern reported. Please add a control condition, such as lower-face masking or distractor sentences matched on length and prosody, or otherwise demonstrate that the 79% versus 8% gap is not an artifact of these confounds.
[§4.4 (semantic appropriateness score), Figs. 13–14] Applying the stated formula Appr_semantic = (P − Pbar)/(P + Pbar + N + B) to the response shares reported for the motion-capture condition in Fig. 13 gives (0.79 − 0.07)/(0.79 + 0.07 + 0.09 + 0.05) = 0.72, whereas Fig. 14 reports 0.77. The displayed distribution and the displayed score are therefore inconsistent under the stated definition. Please reconcile the raw vote counts, the formula, and the reported score; if the discrepancy is due to rounding, state the exact counts.
[§3.2.4 and §4.3] For challenge submissions, the matched and mismatched videos in the dyadic study contain different generated agent motions because the mismatched motion is newly generated from segment B's interlocutor inputs. The pairwise comparison can therefore be won on overall motion naturalness or generation quality rather than on responsiveness to the interlocutor, which would bias the appropriateness score even for a system with good dyadic behaviour. The paper does not report any check that separates generation-quality differences from dyadic responsiveness. Please add such an analysis, for example by correlating matched-versus-mismatched quality ratings or by including a system-generated condition with an unresponsive interlocutor as a sanity check, or soften the conclusion in Sec. 5 that 'systems were not yet able to generate motion that is responsive to the interlocutor.'

minor comments (5)

[§3.2.1] In Section 3.2.1, 'V oice Activity Detection' contains a stray space, and 'V AD' is similarly formatted in Section 3.2.3; please fix the spacing.
[§4.2.1] The symbols \bar{S} and \bar{C} in the appropriateness score formulas are never explicitly defined; please define them the first time they appear.
[References] References [8] and [9] appear to be the same GENEA Challenge 2023 paper; please cite it once.
[§3.2.3] Since the T3 stimulus selection thresholds (30%–50% VAD overlap) are described only as 'determined empirically', a brief sensitivity analysis or a report of the number of candidate windows discarded at each threshold would strengthen reproducibility.
[Abstract and Table 1] The abstract and Table 1 report 'over 23,000 votes' and '869 test-takers'; the table sums to 23,210, so the statement is accurate, but the body text has missing spaces in 'over23,000' and '869test-takers'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the challenge results are measured from collected votes; the only transformed quantity is an explicitly acknowledged rescaling of the prior appropriateness score.

full rationale

The paper does not derive predictions from fitted parameters or define its target quantities in terms of themselves. T1-T4 are pairwise user studies; the reported winrates, Elo ratings, and appropriateness scores are direct aggregations of the 23,000+ collected votes. The updated speech-appropriateness formula is explicitly presented as a linear rescaling of the 2023 GENEA score (Appr_updated = 2(Approld - 0.5)), and the claim that an input-independent system should score zero follows from the exchangeability of matched and mismatched conditions under the definition, not from any fitted value. The semantic appropriateness score is likewise a definition over vote counts, so the mocap ceiling of 77% and the best system's 8% are empirical measurements. The manual curation described in Sec. 3.2 deliberately selects segments with the target qualities; this is a study-design choice that sets the evaluation ceiling, not a parameter fitted to produce a particular system ranking, and the paper explicitly discloses that curation raises the mocap ceiling. Self-citations to [8] and [11] are methodology references to prior published challenge protocols, and the results are not derived from those citations; the leaderboard comparison on BEAT2 provides an external reference point. The skeptical concern that T4's mismatching may measure articulation timing or emphasis rather than semantic expressivity is a validity/confounding threat, not circularity, because the appropriateness score is not defined in terms of, and does not presuppose, the semantic interpretation of the votes. No step in the paper reduces by construction to its own input.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central evaluation rests on the validity of human preference judgments on curated clips and on the assumption that mismatching manipulations isolate each attribute. The paper makes these assumptions explicit in the methodology but does not independently validate them. No free parameters are used in a model derivation; the listed parameters are segment-selection thresholds. No invented entities are introduced.

free parameters (2)

T3 VAD overlap thresholds = lower 30%, upper 50%
Empirically determined from observed dyadic segments to select windows with overlapping speech; directly affects which 46 dyadic segments are evaluated and therefore the mocap ceiling in E3.
Speaking-ratio thresholds for monadic segments = listener at most 10%, main speaker at least 50%
Hand-set selection criteria in Sec 3.2.1; determine the 170 T1 and T2 segments and influence how challenging the evaluation is.

assumptions (4)

domain assumption VAD annotations and transcripts in Seamless Interaction are accurate enough for automated segment selection.
Sec 3.2.1 and Sec 3.2.2 merge VAD segments and align sentence boundaries; errors in these annotations would propagate into the evaluation stimuli.
domain assumption Pairwise preference votes on short rendered clips measure the intended constructs, namely realism, speech alignment, dyadic responsiveness, and semantic expressivity.
All four studies assume construct validity; no validation against objective measures of gesture quality is provided.
domain assumption Mismatched stimuli differ only in the attribute under test.
T2 matches VAD structure but not prosody or voice; T3 mismatched system outputs are newly generated, so generation quality may differ between matched and mismatched videos; T4 uses sentence transcripts as the only difference.
domain assumption Motion capture is an appropriate ceiling for natural gesture quality and appropriateness.
The paper treats filtered mocap segments as the reference ceiling in all four studies; manual filtering can inflate this ceiling relative to uncurated human motion.

how reviews work

0 comments

Cite this review

Pith. "Pith review of The GENEA Challenge 2026: A Large-Scale Disentangled Evaluation of Speech-Driven Gesture Generation on the Seamless Interaction Dataset." pith.science (2026). https://pith.science/paper/J34DT35Y

@misc{pith2026260810839,
  author       = {Pith},
  title        = {Pith review of: The GENEA Challenge 2026: A Large-Scale Disentangled Evaluation of Speech-Driven Gesture Generation on the Seamless Interaction Dataset},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/J34DT35Y}},
  note         = {Machine review of arXiv:2608.10839}
}

read the original abstract

This preprint presents the results of the fourth GENEA Challenge, a large-scale human evaluation of five speech-driven gesture-generation systems trained by participating teams on the Seamless Interaction dataset of dyadic conversations. As in the 2023 GENEA Challenge, we used a disentangled evaluation methodology to assess motion quality and speech alignment without confounding between the two, and performed a dyadic mismatching study to isolate the effect of listening and reacting to the interlocutor. We additionally introduce a new semantic gesture-generation task and a text-mismatching evaluation methodology using the Grounded Gestures subset of the data. In total, we ran four large-scale user studies, collecting over 23,000 votes from 869 test-takers. In the motion-realism study, the dataset's filtered segments had substantially higher motion quality than all challenge submissions (68-95% pairwise winrate). In the speech-alignment study, the motion-capture segments provided a conceptual ceiling at 62% alignment score, with the top submission significantly behind at 32% and the rest only slightly above the 0% expected of an input-independent system. In the dyadic study, motion capture again set the ceiling at 65% appropriateness score, but no submission scored substantially above chance, indicating that the systems could not yet respond to the interlocutor. Finally, the semantic mismatching evaluation found highly expressive gestures in the dataset (test-takers identified the matching transcript 79% of the time), yet almost all submissions failed to generate semantically expressive motion, with the best achieving only an 8% appropriateness score. The collected votes and outputs will be made publicly available at https://genea-workshop.github.io/2026/challenge/ to facilitate reproducibility and further research.

Figures

Figures reproduced from arXiv: 2608.10839 by the authors.

**Figure 2.** 95% confidence intervals for motion realism (Elo) and speech-gesture alignment (appropriateness score) for each condition in the [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗

**Figure 3.** 95% confidence intervals for dyadic alignment versus semantic alignment (appropriateness scores) for each condition, comparing [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

Figures from the paper (11 more)

**Figure 4.** Figure 4: Motion-realism study: breakdown of the votes cast per condition. [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]

**Figure 5.** Figure 5: Motion-realism study: Bradley-Terry Elo ratings for each condition, with 95% confidence intervals obtained via bootstrapping. [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]

**Figure 6.** Figure 6: Motion-realism study: JUICE justifications provided by test-takers. [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]

**Figure 7.** Figure 7: Speech mismatching study: breakdown of the votes cast per condition. [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]

**Figure 8.** Figure 8: Speech mismatching study: appropriateness score for each condition, with 95% confidence intervals obtained via bootstrapping. [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]

**Figure 9.** Figure 9: Speech mismatching study: JUICE justifications provided by test-takers. [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]

**Figure 10.** Figure 10: Dyadic mismatching study: breakdown of the votes cast per condition. [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]

**Figure 11.** Figure 11: Dyadic mismatching study: appropriateness score for each condition, with 95% confidence intervals obtained via bootstrapping. [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]

**Figure 12.** Figure 12: Dyadic mismatching study: JUICE justifications provided by test-takers. [PITH_FULL_IMAGE:figures/full_fig_p014_12.png]

**Figure 13.** Figure 13: Semantic mismatching study: breakdown of the votes cast per condition. [PITH_FULL_IMAGE:figures/full_fig_p014_13.png]

**Figure 14.** Figure 14: Semantic mismatching study: appropriateness score for each condition, with 95% confidence intervals obtained via bootstrap [PITH_FULL_IMAGE:figures/full_fig_p015_14.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

11 extracted references · 10 canonical work pages

[1]

Seamless interaction: Dyadic audiovisual motion modeling and large-scale dataset.arXiv preprint arXiv:2506.22554,

Vasu Agrawal, Akinniyi Akinyemi, Kathryn Alvero, Morteza Behrooz, Julia Buffalini, Fabio Maria Carlucci, Joy Chen, Junming Chen, Zhang Chen, Shiyang Cheng, et al. Seamless interaction: Dyadic audiovisual motion modeling and large-scale dataset.arXiv preprint arXiv:2506.22554,

arXiv
[2]

UM-FERI approach to GENEA challenge 2026, 2026

Karlo Crnek. UM-FERI approach to GENEA challenge 2026, 2026. Non-archival paper, to appear in the Interactive Social Agents Workshop at ECCV 2026. 2

work page 2026
[3]

DyaSync: Disentangling speech identity and activity for dyadic co- speech gesture generation, 2026

Fengyi Fang, Ye Lu, Qian Bao, and Xudong Liu. DyaSync: Disentangling speech identity and activity for dyadic co- speech gesture generation, 2026. Non-archival paper, to ap- pear in the Interactive Social Agents Workshop at ECCV

work page 2026
[4]

Factorizing text-to-video generation by explicit image conditioning

Rohit Girdhar, Mannat Singh, Andrew Brown, Quentin Du- val, Samaneh Azadi, Sai Saketh Rambhatla, Akbar Shah, Xi Yin, Devi Parikh, and Ishan Misra. Factorizing text-to-video generation by explicit image conditioning. pages 205–224,

work page
[5]

Johsac Isbac Gomez Sanchez and Paula D. P. Costa. A uni- fied flow-matching DiT with direct continuous decoding for co-speech gesture generation. InProceedings of the Euro- pean Conference on Computer Vision (ECCV) Workshops,

work page
[6]

GestFlow: A small flow-matching system for the GENEA challenge 2026

Jos ´e Guillen. GestFlow: A small flow-matching system for the GENEA challenge 2026. InProceedings of the European Conference on Computer Vision (ECCV) Workshops, 2026. To appear. 2

work page 2026
[7]

A large, crowdsourced eval- uation of gesture generation systems on common data: The genea challenge 2020

Taras Kucherenko, Patrik Jonell, Youngwoo Yoon, Pieter Wolfert, and Gustav Eje Henter. A large, crowdsourced eval- uation of gesture generation systems on common data: The genea challenge 2020. In26th international conference on intelligent user interfaces, pages 11–21, 2021. 1, 2

work page 2020
[8]

The genea challenge 2023: A large-scale evaluation of gesture generation models in monadic and dyadic settings

Taras Kucherenko, Rajmund Nagy, Youngwoo Yoon, Jieyeon Woo, Teodor Nikolov, Mihail Tsakov, and Gus- tav Eje Henter. The genea challenge 2023: A large-scale evaluation of gesture generation models in monadic and dyadic settings. InProceedings of the 25th International Conference on Multimodal Interaction, page 792–801, New York, NY , USA, 2023. Association...

work page 2023

Show all 11 references

[9]

The GENEA Challenge 2023: A large- scale evaluation of gesture generation models in monadic and dyadic settings

Taras Kucherenko, Rajmund Nagy, Youngwoo Yoon, Jieyeon Woo, Teodor Nikolov, Mihail Tsakov, and Gus- tav Eje Henter. The GENEA Challenge 2023: A large- scale evaluation of gesture generation models in monadic and dyadic settings. InProceedings of the International Con- ference ...

2023
[10]

Evaluating gesture generation in a large-scale open chal- lenge: The GENEA Challenge 2022.ACM Transactions on Graphics (TOG), 2024

Taras Kucherenko, Pieter Wolfert, Youngwoo Yoon, Carla Viegas, Teodor Nikolov, Mihail Tsakov, and Gustav Eje Hen- ter. Evaluating gesture generation in a large-scale open chal- lenge: The GENEA Challenge 2022.ACM Transactions on Graphics (TOG), 2024. 1, 2

2022
[11]

Towards reli- able human evaluations in gesture generation: Insights from a community-driven state-of-the-art benchmark

Rajmund Nagy, Hendric V oss, Thanh Hoang-Minh, Mihail Tsakov, Teodor Nikolov, Zeyi Zhang, Tenglong Ao, Sicheng Yang, Shaoli Huang, Yongkang Cheng, et al. Towards reli- able human evaluations in gesture generation: Insights from a community-driven state-of-the-art benchmark. In...

2026

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

Seamless interaction: Dyadic audiovisual motion modeling and large-scale dataset.arXiv preprint arXiv:2506.22554,

Vasu Agrawal, Akinniyi Akinyemi, Kathryn Alvero, Morteza Behrooz, Julia Buffalini, Fabio Maria Carlucci, Joy Chen, Junming Chen, Zhang Chen, Shiyang Cheng, et al. Seamless interaction: Dyadic audiovisual motion modeling and large-scale dataset.arXiv preprint arXiv:2506.22554,

arXiv

[2] [2]

UM-FERI approach to GENEA challenge 2026, 2026

Karlo Crnek. UM-FERI approach to GENEA challenge 2026, 2026. Non-archival paper, to appear in the Interactive Social Agents Workshop at ECCV 2026. 2

work page 2026

[3] [3]

DyaSync: Disentangling speech identity and activity for dyadic co- speech gesture generation, 2026

Fengyi Fang, Ye Lu, Qian Bao, and Xudong Liu. DyaSync: Disentangling speech identity and activity for dyadic co- speech gesture generation, 2026. Non-archival paper, to ap- pear in the Interactive Social Agents Workshop at ECCV

work page 2026

[4] [4]

Factorizing text-to-video generation by explicit image conditioning

Rohit Girdhar, Mannat Singh, Andrew Brown, Quentin Du- val, Samaneh Azadi, Sai Saketh Rambhatla, Akbar Shah, Xi Yin, Devi Parikh, and Ishan Misra. Factorizing text-to-video generation by explicit image conditioning. pages 205–224,

work page

[5] [5]

Johsac Isbac Gomez Sanchez and Paula D. P. Costa. A uni- fied flow-matching DiT with direct continuous decoding for co-speech gesture generation. InProceedings of the Euro- pean Conference on Computer Vision (ECCV) Workshops,

work page

[6] [6]

GestFlow: A small flow-matching system for the GENEA challenge 2026

Jos ´e Guillen. GestFlow: A small flow-matching system for the GENEA challenge 2026. InProceedings of the European Conference on Computer Vision (ECCV) Workshops, 2026. To appear. 2

work page 2026

[7] [7]

A large, crowdsourced eval- uation of gesture generation systems on common data: The genea challenge 2020

Taras Kucherenko, Patrik Jonell, Youngwoo Yoon, Pieter Wolfert, and Gustav Eje Henter. A large, crowdsourced eval- uation of gesture generation systems on common data: The genea challenge 2020. In26th international conference on intelligent user interfaces, pages 11–21, 2021. 1, 2

work page 2020

[8] [8]

The genea challenge 2023: A large-scale evaluation of gesture generation models in monadic and dyadic settings

Taras Kucherenko, Rajmund Nagy, Youngwoo Yoon, Jieyeon Woo, Teodor Nikolov, Mihail Tsakov, and Gus- tav Eje Henter. The genea challenge 2023: A large-scale evaluation of gesture generation models in monadic and dyadic settings. InProceedings of the 25th International Conference on Multimodal Interaction, page 792–801, New York, NY , USA, 2023. Association...

work page 2023