Pith. sign in

REVIEW 4 major objections 8 minor 122 references

NTIRE 2025 XGC Quality Assessment Challenge: Methods and Results

T0 review · 4 major / 8 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper reports a three-track competition and claims that all 19 submitted quality-assessment models beat their baselines, with the largest gain on AI-generated video.

desk verdict Useful challenge leaderboard for VQA, but the abstract's 'all teams beat baseline' claim is internally contradicted by the paper's own track-3 counts. read the letter →

arxiv 2506.02875 v1 pith:A4QJEMUB submitted 2025-06-03 cs.CV

Xiaohong Liu , Xiongkuo Min , Qiang Hu , Xiaoyun Zhang , Jie Guo , Guangtao Zhai , Shushi Wang , Yingjie Zhou
show 87 more authors
Lu Liu Jingxin Li Liu Yang Farong Wen Li Xu Yanwei Jiang Xilei Zhu Chunyi Li Zicheng Zhang Huiyu Duan Xiele Wu Yixuan Gao Yuqin Cao Jun Jia Wei Sun Jiezhang Cao Radu Timofte Baojun Li Jiamian Huang Dan Luo Tao Liu Weixia Zhang Bingkun Zheng Junlin Chen Ruikai Zhou Meiya Chen Yu Wang Hao Jiang Xiantao Li Yuxiang Jiang Jun Tang Yimeng Zhao Bo Hu Zelu Qi Chaoyang Zhang Fei Zhao Ping Shi Lingzhi Fu Heng Cong Shuai He Rongyu Zhang Jiarong He Zongyao Hu Wei Luo Zihao Yu Fengbin Guan Yiting Lu Xin Li Zhibo Chen Mengjing Su Yi Wang Tuo Chen Chunxiao Li Shuaiyu Zhao Jiaxin Wen Chuyi Lin Sitong Liu Ningxin Chu Jing Wan Yu Zhou Baoying Chen Jishen Zeng Jiarui Liu Xianjin Liu Xin Chen Lanzhi Zhou Hangyu Li You Han Bibo Xiang Zhenjie Liu Jianzhang Lu Jialin Gui Renjie Lu Shangfei Wang Donghao Zhou Jingyu Lin Quanjian Song Jiancheng Huang Yufeng Yang Changwei Wang Shupeng Zhong Yang Yang Lihuo He Jia Liu Yuting Xing Tida Fang Yuchun Jin
This is my paper · ORCID
classification cs.CV
keywords videoqualityassessmentuser-generatedAI-generatedtalkingheadNTIREchallengereportmeanopinionscoreSpearmanrankcorrelationmultimodal
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper reports the NTIRE 2025 XGC Quality Assessment Challenge, a three-track competition to predict how humans perceive the quality of user-generated videos, AI-generated videos, and talking heads. Its central claim is that all 19 submitted models outperformed the designated baselines: track winners reached main scores of 0.8731, 0.6645, and 0.8244 against baseline scores of 0.7789, 0.5161, and 0.7862 respectively. A sympathetic reader should take away that content-type-specific quality assessment is a workable and substantially better strategy than applying a generic video-quality model across domains, especially for AI-generated video where the best baseline sits at 0.5161. The challenge also contributes three public datasets, FineVD-GC, Q-Eval-Video, and THQA-NTIRE, and a documented evaluation protocol centered on the average of Spearman and Pearson correlations.

What carries the argument

The load-bearing mechanism is the challenge's evaluation metric, the main score, defined as the average of the absolute Spearman rank correlation and the Pearson linear correlation after third-order polynomial regression: Main Score = $(|SRCC| + |PLCC|)/2$. The other half of the machinery is the three datasets used as test beds: FineVD-GC (6,284 user-generated videos with mean opinion scores from 120 subjects), Q-Eval-Video (34,029 AI-generated videos from 11 text-to-video models, rated with the Sample and Scrutinize strategy), and THQA-NTIRE (12,247 talking heads combining THQA and THQA-3D, all with audio). The rankings in the result tables and the claim that every team beats the baseline are computed entirely from this metric on these held-out test splits.

What would settle it

Re-annotate a random sample of videos from each test split with a fresh panel of subjects and recompute the main scores; if the resampled mean opinion scores put several submitted models within noise of the baselines or move the 0.0008 first-versus-second gap in the talking head track, the central claim loses its force. A cheaper check is to compute confidence intervals or bootstrap resampling for the SRCC and PLCC differences in the three result tables, which the paper does not report.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that specialized, content-aware quality assessment models beat general-purpose baselines on held-out test splits in all three tracks, with the size of the win varying by content type. In the user-generated video track the winning model scored 0.8731 against the FastVQA baseline's 0.7789; in the AI-generated video track the winner scored 0.6645 against the best baseline T2VQA's 0.5161; and in the talking head track the winner scored 0.8244 against SimpleVQA's 0.7862, with the top two finishers separated by only 0.0008. The winning methods are respectively a multimodal large language model fine-tuned with low-rank adaptation and spatial window sampling, a temporal pyramid sampling approach that captures motion smoothness and content authenticity at multiple frame rates, and an audio-visual fusion network combining spatial features from still frames, temporal features from video clips, and audio time-frequency features.

Load-bearing premise

The load-bearing premise is that the mean opinion scores in FineVD-GC, Q-Eval-Video, and THQA-NTIRE are accurate and that the random 4:1:1 test splits represent their domains, because every ranking and the claim that all teams beat the baselines rest on those labels, and the paper reports no label-noise or subject-agreement statistics.

Editorial extensions

If this is right

  • Training a quality model on a specific content type, UGC, AIGV, or talking heads, improves correlation with human opinion over off-the-shelf baselines across all three tracks.
  • The largest headroom is in AI-generated video, where the best submitted model at 0.6645 beats the best baseline T2VQA at 0.5161 by a wide margin, indicating that generic models fall shortest on this content type.
  • Fine-grained UGC quality is predictable: the winning track-1 model scored separate dimensions, color, noise, artifact, blur, and temporal, and still led overall, suggesting multidimensional assessment does not sacrifice accuracy.
  • For talking heads, audio information is usable: the winning architecture fuses visual features with time-frequency audio features, and every submitted model outscored the visual-only baseline SimpleVQA.
  • The three datasets and the two-phase leaderboard protocol provide a reusable public benchmark for future quality assessment research on these content types.

Reading between the lines

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

  • Because the test sets come from the same datasets as the training sets, the reported gains are in-distribution; whether specialized models beat generic baselines on unseen UGC or AIGV corpora remains untested in this paper.
  • The 0.0008 point gap between first and second place in the talking-head track is smaller than typical MOS noise, so the top of that leaderboard should probably be read as a tie until significance or confidence intervals are reported.
  • The low absolute main score in the AI-generated video track, 0.6645, suggests that predicting human opinion for AIGVs is still an open problem, and a natural next application is using these scores to select among text-to-video generation models.
  • A testable extension is whether the winning MLLM-based UGC method transfers to talking heads or AI-generated videos without the temporal or audio modules the other track winners found necessary.
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

4 major / 8 minor

Summary. The paper reports the organization and results of the NTIRE 2025 XGC Quality Assessment Challenge, which comprises three tracks: user-generated video, AI-generated video, and talking-head quality assessment. It describes the benchmark datasets (FineVD-GC, Q-Eval-Video, THQA-NTIRE), the evaluation protocol in which the main score is the average of |SRCC| and |PLCC| after third-order polynomial regression, and the final test-phase rankings of the participating teams together with baseline comparisons. The winners of the three tracks are described in Section 5. The central claim of the paper, stated in the abstract, is that every participating team in every track proposed a method that outperforms the baseline.

Significance. If the reported measurements are reliable, the challenge provides a useful snapshot of specialized VQA models for three content types, with the best models substantially outperforming generic baselines (e.g., 0.8731 vs 0.7789 in Track 1 and 0.6645 vs 0.5161 in Track 2). The paper's strengths are the clear and reproducible scoring formula, the use of held-out test-phase predictions for ranking, and the collection of per-team fact sheets described in the supplementary material. The main score and all table entries are internally consistent row-by-row, and the reported test measurements are not fitted to the leaderboard. However, the significance is currently limited by unresolved inconsistencies in the number of valid teams, a missing method description for the Track 2 winner, and the complete absence of uncertainty quantification or MOS reliability information, which prevents the reader from fully trusting the very tight ranking margins.

major comments (4)
  1. [Section 4.2 / Table 3 / Section 1 / Abstract] The number of valid talking-head teams is reported inconsistently. Section 1 states that there are 9 valid participating teams in the talking head track, while the Abstract, Section 4, and Table 3 all list 8 teams. Moreover, Section 4.2 says '7 teams achieve a main score higher than baseline,' but Table 3 lists 8 teams, all with main scores above the SimpleVQA baseline of 0.7862; the lowest listed score, XIDIAN-VQA at 0.7872, is still above the baseline. The abstract's central claim that every team in every track outperforms the baseline is therefore contradicted by one or more of these statements, and no single reading makes all of them true. If 9 teams were valid, Table 3 is missing a team; if only 7 teams beat the baseline, the 0.7872 value must be wrong. The counts and the table/prose must be reconciled.
  2. [Section 5.2] The description of the Track 2 winning method is truncated and corrupted. After the sentence fragment ending with 'to ad-', the text is replaced by a long block of '/uni000...' tokens, and the Figure 3 caption then appears at the end of this section. No substantive description of the temporal pyramid sampling method is actually provided. Since the paper is titled 'Methods and Results' and Section 5 is meant to describe the winning methods, the missing technical content for the AI-generated-video winner is a substantive gap, not a mere typo.
  3. [Section 3.2 / References [22] and [45]] The FineVD-GC dataset is cited inconsistently. Section 3.2 cites [22] for FineVD-GC, but reference [22] is the FineVQ paper, while the Abstract and Section 1 cite [45] for the same dataset, and reference [45] is the AIGIQA-20k paper. Neither citation clearly identifies the FineVD-GC companion paper. Since the datasets are the foundation of the entire challenge and all ranking claims, the correct provenance must be stated precisely.
  4. [Section 3.2 / Section 4.2 / Table 3] The paper provides no evidence that the Mean Opinion Scores are reliable enough to support the reported ranking differences. No subject agreement, label-noise, or annotation-reliability statistics are reported for any of the three datasets, and the representativeness of the test splits is asserted without supporting analysis. This matters concretely: in Table 3, the top two teams differ by only 0.0008 in main score, and the claim that the 8th team beats the baseline rests on a 0.0010 margin (0.7872 vs 0.7862). The paper should report confidence intervals or bootstrap significance tests for the SRCC/PLCC values, or at minimum report MOS reliability statistics, before making universal 'outperforms the baseline' claims.
minor comments (8)
  1. [Section 3.3] The evaluation-protocol paragraph says 'In both tracks' even though the paper describes three tracks; it should say 'In all three tracks' or similar.
  2. [Section 3.3] The phrase 'Person Linear Correlation Coefficient' should be 'Pearson Linear Correlation Coefficient'.
  3. [Section 3.3 / Figures 1 and 2 captions] Section 3.3 specifies a third-order polynomial nonlinear regression before computing PLCC, but the captions of Figures 1 and 2 say the curves are obtained by a 'four-order polynomial nonlinear fitting'. Please clarify which fitting order is used and make the text consistent.
  4. [Table 1] Table 1 lists 'Overall' as if it were a quality dimension alongside Color, Noise, Artifact, Blur, and Temporal, while also reporting a separate 'Main Score'. The meaning of the 'Overall' column and its relation to the main score should be explained.
  5. [References] References [68] and [69] appear to be the same SimpleVQA paper, and references [83] and [84] appear to be the same Fast-VQA paper; these should be merged or cross-referenced to avoid duplicate citations.
  6. [Section 4] There is a typo 'Tabel 3' that should read 'Table 3'.
  7. [Section 3.1 / 3.2] Section 3.1 refers to 'THQA[117, 118]' while Section 3.2 uses 'THQA-NTIRE'; the dataset name should be used consistently and defined at first use.
  8. [Figure 3] The Figure 3 caption says it shows talking-head track performance, but the caption appears at the end of Section 5.2 (the AI-generated video section) rather than in the Section 4 results for Track 3; the placement should be corrected.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: challenge scores are held-out test measurements against human MOS; self-cited datasets are benchmark infrastructure, not derivation inputs.

full rationale

No circular step can be exhibited in this challenge report. The abstract's claim that every submitted method outperforms the baseline is an empirical statement about test-set measurements: Section 3.2 describes a 4:1:1 random split into training/validation/testing, Section 3.4 states that MOS for the testing set are withheld during the challenge, and Tables 1-3 report SRCC/PLCC main scores computed via Eq. (1) against human MOS. These correlations are held-out measurements, not quantities fitted from the same data and renamed as predictions. The datasets (FineVD-GC [22], Q-Eval-Video [101], THQA [117,118]) and several baselines (Q-Eval-Score [101], T2VQA [42], FastVQA [84], DOVER [86], SimpleVQA [69]) are products of the organizing group, so the benchmark is self-anchored; however, the MOS are human subjective annotations external to the submitted models, and no score in Tables 1-3 is defined in terms of the baseline or the challenge outcome. Internal problems are present but are correctness issues, not circularity: Section 1 says 9 valid talking-head teams while the abstract, Section 4 and Table 3 say 8; Section 4.2 says 7 teams beat the talking-head baseline while Table 3 lists all 8 teams above SimpleVQA; Section 5.2 contains a corrupted unicode block; and Section 3.2 reports no inter-subject agreement statistics for the MOS. None of these make a 'prediction' equivalent to its input by construction, so the circularity score is 0.

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

This is a challenge report, so the ledger is lighter than for a theory paper. The central result is a set of measured correlations from Section 4; what the reader pays for upstream is the ground truth labels (three datasets produced by the organizing group), the evaluation protocol (standard VQA practice), and the baseline configurations (undisclosed). The third-order polynomial regression in Section 3.3 is a fitting step embedded inside the reported main score.

free parameters (3)
  • Third-order polynomial regression coefficients = per-method coefficients fitted on test predictions
    Section 3.3: PLCC is computed after third-order polynomial nonlinear regression of each submission's predictions. This is standard VQA calibration, but it is a data-dependent fitting step inside the reported main score and is applied to all methods.
  • FineVD-GC MOS aggregation settings = 120 subjects, normalized and averaged
    Section 3.2: the track 1 ground truth rests on a subjective study described in the cited dataset paper, with screening and normalization details not repeated here; the ranking inherits those choices.
  • Q-Eval-Video Sample and Scrutinize settings = not specified in this paper
    Section 3.2: the track 2 ground truth depends on the Sample and Scrutinize annotation strategy from [101]; its thresholds and subject counts are not given here.
assumptions (3)
  • domain assumption The Mean Opinion Scores in FineVD-GC, Q-Eval-Video, and THQA-NTIRE accurately reflect human perceptual quality.
    Section 3.2 describes dataset construction, but the label accuracy is inherited from the companion papers [22], [101], [117, 118]; the entire ranking depends on these labels being correct.
  • domain assumption SRCC and PLCC after third-order polynomial calibration are a valid measure of quality assessment performance.
    Section 3.3 defines the main score as (|SRCC|+|PLCC|)/2; the field-standard protocol is assumed rather than justified.
  • domain assumption The test splits are representative of user-generated video, AI-generated video, and talking head content.
    Section 3.2 gives a 4:1:1 random split for FineVD-GC; the split procedures for Q-Eval-Video and THQA-NTIRE are not described, so representativeness is assumed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of NTIRE 2025 XGC Quality Assessment Challenge: Methods and Results." pith.science (2026). https://pith.science/paper/A4QJEMUB

@misc{pith2026250602875,
  author       = {Pith},
  title        = {Pith review of: NTIRE 2025 XGC Quality Assessment Challenge: Methods and Results},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/A4QJEMUB}},
  note         = {Machine review of arXiv:2506.02875}
}
read the original abstract

This paper reports on the NTIRE 2025 XGC Quality Assessment Challenge, which will be held in conjunction with the New Trends in Image Restoration and Enhancement Workshop (NTIRE) at CVPR 2025. This challenge is to address a major challenge in the field of video and talking head processing. The challenge is divided into three tracks, including user generated video, AI generated video and talking head. The user-generated video track uses the FineVD-GC, which contains 6,284 user generated videos. The user-generated video track has a total of 125 registered participants. A total of 242 submissions are received in the development phase, and 136 submissions are received in the test phase. Finally, 5 participating teams submitted their models and fact sheets. The AI generated video track uses the Q-Eval-Video, which contains 34,029 AI-Generated Videos (AIGVs) generated by 11 popular Text-to-Video (T2V) models. A total of 133 participants have registered in this track. A total of 396 submissions are received in the development phase, and 226 submissions are received in the test phase. Finally, 6 participating teams submitted their models and fact sheets. The talking head track uses the THQA-NTIRE, which contains 12,247 2D and 3D talking heads. A total of 89 participants have registered in this track. A total of 225 submissions are received in the development phase, and 118 submissions are received in the test phase. Finally, 8 participating teams submitted their models and fact sheets. Each participating team in every track has proposed a method that outperforms the baseline, which has contributed to the development of fields in three tracks.

Figures

Figures reproduced from arXiv: 2506.02875 by the authors.

Figure 1
Figure 1. Scatter plots of the predicted scores vs. MOSs in the user-generated video track. The curves are obtained by a four-order [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Scatter plots of the predicted scores vs. MOSs in the AI generated video track. The curves are obtained by a four-order polynomial [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. The performance of methods proposed by different teams in Talking head track. [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

122 extracted references · 52 canonical work pages

  1. [22]

    Finevq: Fine-grained user gen- erated content video quality assessment.arXiv preprint arXiv:2412.19238, 2024

    Huiyu Duan, Qiang Hu, Jiarui Wang, Liu Yang, Zitong Xu, Lu Liu, Xiongkuo Min, Chunlei Cai, Tianxiao Ye, Xiaoyun Zhang, et al. Finevq: Fine-grained user gen- erated content video quality assessment.arXiv preprint arXiv:2412.19238, 2024. 2, 4

    arXiv 2024

  2. [45]

    Aigiqa-20k: A large database for ai-generated image quality assessment

    Chunyi Li, Tengchuan Kou, Yixuan Gao, Yuqin Cao, Wei Sun, Zicheng Zhang, Yingjie Zhou, Zhichao Zhang, Haon- ing Wu, Weixia Zhang, Xiaohong Liu, Xiongkuo Min, and Guangtao Zhai. Aigiqa-20k: A large database for ai-generated image quality assessment. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2024. 2

    2024

  3. [1]

    Inhttps://runwayml.com/research/introducing-gen-3- alpha, 2024

    Introducing gen-3 alpha: A new frontier for video genera- tion. Inhttps://runwayml.com/research/introducing-gen-3- alpha, 2024. 4

    2024

  4. [2]

    Audio-visual face reen- actment

    Madhav Agarwal, Rudrabha Mukhopadhyay, Vinay P Namboodiri, and CV Jawahar. Audio-visual face reen- actment. InProceedings of the IEEE/CVF winter confer- ence on applications of computer vision, pages 5178–5187,

  5. [3]

    Dream machine: Ai video generator

    Luma AI. Dream machine: Ai video generator. In https://lumalabs.ai/dream-machine, 2024. 4

    2024

  6. [4]

    Pixverse: Ai video creation platform

    PixVerse AI. Pixverse: Ai video creation platform. In https://pixverse.ai/, 2024. 4

    2024

  7. [5]

    Vidu AI Team. Vidu ai. Inhttps://www.vidu.studio/zh,

  8. [6]

    Stable video diffusion: Scaling latent video diffusion models to large datasets

    Andreas Blattmann, Tim Dockhorn, Sumith Kulal, Daniel Mendelevitch, Maciej Kilian, Dominik Lorenz, Yam Levi, Zion English, Vikram V oleti, and Adam Letts. Stable video diffusion: Scaling latent video diffusion models to large datasets. InarXiv preprint:2311.15127, 2023. 4

    arXiv 2023

Show all 122 references
  1. [7]

    Finding directions in gan’s latent space for neural face reenactment.arXiv preprint arXiv:2202.00046,

    Stella Bounareli, Vasileios Argyriou, and Georgios Tz- imiropoulos. Finding directions in gan’s latent space for neural face reenactment.arXiv preprint arXiv:2202.00046,

    arXiv

  2. [8]

    Stylemask: Disentangling the style space of stylegan2 for neural face reenactment

    Stella Bounareli, Christos Tzelepis, Vasileios Argyriou, Ioannis Patras, and Georgios Tzimiropoulos. Stylemask: Disentangling the style space of stylegan2 for neural face reenactment. In2023 IEEE 17th international conference on automatic face and gesture recognition (FG), pag...

    2023

  3. [9]

    Dreamina

    Dreamina by CapCut. Dreamina. In https://dreamina.capcut.com/, 2023. 4

    2023

  4. [10]

    Internlm2 technical report, 2024

    Zheng Cai, Maosong Cao, Haojiong Chen, Kai Chen, Keyu Chen, Xin Chen, Xun Chen, Zehui Chen, Zhi Chen, Pei Chu, Xiaoyi Dong, Haodong Duan, Qi Fan, Zhaoye Fei, Yang Gao, Jiaye Ge, Chenya Gu, Yuzhe Gu, Tao Gui, Aijia Guo, Qipeng Guo, Conghui He, Yingfan Hu, Ting Huang, Tao Jiang,...

    2024

  5. [11]

    A no-reference quality assessment metric for dynamic 3d digital human.Displays, 80:102540, 2023

    Shi Chen, Zicheng Zhang, Yingjie Zhou, Wei Sun, and Xiongkuo Min. A no-reference quality assessment metric for dynamic 3d digital human.Displays, 80:102540, 2023. 3

    2023

  6. [12]

    NTIRE 2025 challenge on image super-resolution (×4): Methods and results

    Zheng Chen, Kai Liu, Jue Gong, Jingkai Wang, Lei Sun, Zongwei Wu, Radu Timofte, Yulun Zhang, et al. NTIRE 2025 challenge on image super-resolution (×4): Methods and results. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Work- shops, 2025. 2

    2025

  7. [13]

    NTIRE 2025 challenge on real-world face restoration: Methods and results

    Zheng Chen, Jingkai Wang, Kai Liu, Jue Gong, Lei Sun, Zongwei Wu, Radu Timofte, Yulun Zhang, et al. NTIRE 2025 challenge on real-world face restoration: Methods and results. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Work- shops, 2025. 2

    2025

  8. [14]

    Videoretalking: Audio-based lip synchro- nization for talking head video editing in the wild

    Kun Cheng et al. Videoretalking: Audio-based lip synchro- nization for talking head video editing in the wild. InSIG- GRAPH Asia, 2022. 3

    2022

  9. [15]

    Measuring the quality of text-to-video model outputs: Metrics and dataset.arXiv preprint arXiv:2309.08009, 2023

    Iya Chivileva, Philip Lynch, Tomas E Ward, and Alan F Smeaton. Measuring the quality of text-to-video model outputs: Metrics and dataset.arXiv preprint arXiv:2309.08009, 2023. 2

    2023 arXiv

  10. [16]

    NTIRE 2025 challenge on raw image restoration and super-resolution

    Marcos Conde, Radu Timofte, et al. NTIRE 2025 challenge on raw image restoration and super-resolution. InProceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2025. 2

    2025

  11. [17]

    Raw image reconstruc- tion from RGB on smartphones

    Marcos Conde, Radu Timofte, et al. Raw image reconstruc- tion from RGB on smartphones. NTIRE 2025 challenge re- port. InProceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR) Workshops,

    2025

  12. [18]

    Francesca De Simone, Marco Tagliasacchi, Matteo Nac- cari, Stefano Tubaro, and Touradj Ebrahimi. A h. 264/avc video database for the evaluation of quality metrics. In 2010 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 2430–2433. IEEE,

    2010

  13. [19]

    Light-vqa: A multi- dimensional quality assessment model for low-light video enhancement

    Yunlong Dong, Xiaohong Liu, Yixuan Gao, Xunchu Zhou, Tao Tan, and Guangtao Zhai. Light-vqa: A multi- dimensional quality assessment model for low-light video enhancement. InProceedings of the 31st ACM Interna- tional Conference on Multimedia, pages 1088–1097, 2023. 2

    2023

  14. [20]

    Light-vqa: A multi- dimensional quality assessment model for low-light video enhancement

    Yunlong Dong, Xiaohong Liu, Yixuan Gao, Xunchu Zhou, Tao Tan, and Guangtao Zhai. Light-vqa: A multi- dimensional quality assessment model for low-light video enhancement. InProceedings of the 31st ACM Interna- tional Conference on Multimedia, 2023. 3

    2023

  15. [21]

    Headgan: One-shot neural head synthesis and editing

    Michail Christos Doukas, Stefanos Zafeiriou, and Viktoriia Sharmanska. Headgan: One-shot neural head synthesis and editing. InProceedings of the IEEE/CVF International con- ference on Computer Vision, pages 14398–14407, 2021. 3

    2021

  16. [23]

    NTIRE 2025 challenge on night photography rendering

    Egor Ershov, Sergey Korchagin, Alexei Khalin, Artyom Panshin, Arseniy Terekhin, Ekaterina Zaychenkova, Georgiy Lobarev, Vsevolod Plokhotnyuk, Denis Abramov, Elisey Zhdanov, Sofia Dorogova, Yasin Mamedov, Nikola Banic, Georgii Perevozchikov, Radu Timofte, et al. NTIRE 2025 chal...

    2025

  17. [24]

    Slowfast networks for video recognition

    Christoph Feichtenhofer, Haoqi Fan, Jitendra Malik, and Kaiming He. Slowfast networks for video recognition. In Proceedings of the IEEE/CVF international conference on computer vision, pages 6202–6211, 2019. 8

    2019

  18. [25]

    NTIRE 2025 challenge on cross-domain few-shot object detection: Methods and results

    Yuqian Fu, Xingyu Qiu, Bin Ren Yanwei Fu, Radu Timofte, Nicu Sebe, Ming-Hsuan Yang, Luc Van Gool, et al. NTIRE 2025 challenge on cross-domain few-shot object detection: Methods and results. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CV...

    2025

  19. [26]

    VDPVE: VQA dataset for perceptual video enhance- ment.arXiv preprint arXiv:2303.09290, 2023

    Yixuan Gao, Yuqin Cao, Tengchuan Kou, Wei Sun, Yun- long Dong, Xiaohong Liu, Xiongkuo Min, and Guangtao Zhai. VDPVE: VQA dataset for perceptual video enhance- ment.arXiv preprint arXiv:2303.09290, 2023. 2

    2023 arXiv

  20. [27]

    High-fidelity and freely controllable talking head video generation

    Yue Gao, Yuan Zhou, Jinglu Wang, Xiao Li, Xiang Ming, and Yan Lu. High-fidelity and freely controllable talking head video generation. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 5609–5619, 2023. 3

    2023

  21. [28]

    Gen-2: Generate novel videos with text, images or video clips

    Anastasis Germanidis. Gen-2: Generate novel videos with text, images or video clips. In https://runwayml.com/research/gen-2, 2023. 4

    2023

  22. [29]

    Deepti Ghadiyaram, Janice Pan, Alan C Bovik, Anush Kr- ishna Moorthy, Prasanjit Panda, and Kai-Chieh Yang. In- capture mobile video distortions: A study of subjective behavior and objective algorithms.IEEE Transactions on Circuits and Systems for Video Technology (TCSVT), 28(9...

    2017

  23. [30]

    An efficient ophthalmic dis- ease qa system integrated with knowledge graphs and digi- tal humans

    Shihan Guo, Jiachen Guo, Han Wang, Haibo Wang, Xiaol- ing Huang, and Lin Zhang. An efficient ophthalmic dis- ease qa system integrated with knowledge graphs and digi- tal humans. In2024 7th International Conference on Infor- mation Communication and Signal Processing (ICICSP),...

    2024

  24. [31]

    Marionette: Few-shot face reenact- ment preserving identity of unseen targets

    Sungjoo Ha, Martin Kersner, Beomsu Kim, Seokjun Seo, and Dongyoung Kim. Marionette: Few-shot face reenact- ment preserving identity of unseen targets. InProceed- ings of the AAAI conference on artificial intelligence, vol- ume 34, pages 10893–10900, 2020. 3

    2020

  25. [32]

    Full-reference and no- reference quality assessment for video frame interpolation

    Jinliang Han, Xiongkuo Min, Jun Jia, Yixuan Gao, Xi- aohong Liu, and Guangtao Zhai. Full-reference and no- reference quality assessment for video frame interpolation. InIEEE Transactions on Circuits and Systems for Video Technology, 2025. 2

    2025

  26. [33]

    NTIRE 2025 challenge on text to image generation model qual- ity assessment

    Shuhao Han, Haotian Fan, Fangyuan Kong, Wenjie Liao, Chunle Guo, Chongyi Li, Radu Timofte, et al. NTIRE 2025 challenge on text to image generation model qual- ity assessment. InProceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR) Workshops...

    2025

  27. [34]

    Depth-aware generative adversarial network for talking head video generation

    Fa-Ting Hong, Longhao Zhang, Li Shen, and Dan Xu. Depth-aware generative adversarial network for talking head video generation. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 3397–3406, 2022. 3

    2022

  28. [35]

    The konstanz natural video database (konvid-1k)

    Vlad Hosu, Franz Hahn, Mohsen Jenadeleh, Hanhe Lin, Hui Men, Tam´as Szir´anyi, Shujun Li, and Dietmar Saupe. The konstanz natural video database (konvid-1k). InPro- ceedings of the 2017 Ninth International Conference on Quality of Multimedia Experience (QoMEX), pages 1–6. IEEE...

    2017

  29. [36]

    Varfvv: View- adaptive real-time interactive free-view video streaming with edge computing.IEEE Journal on Selected Areas in Communications, pages 1–1, 2025

    Qiang Hu, Qihan He, Houqiang Zhong, Guo Lu, Xiaoyun Zhang, Guangtao Zhai, and Yanfeng Wang. Varfvv: View- adaptive real-time interactive free-view video streaming with edge computing.IEEE Journal on Selected Areas in Communications, pages 1–1, 2025. 2

    2025

  30. [37]

    VBench: Comprehensive benchmark suite for video generative mod- els

    Ziqi Huang, Yinan He, Jiashuo Yu, Fan Zhang, Chenyang Si, Yuming Jiang, Yuanhan Zhang, Tianxing Wu, Qingyang Jin, Nattapol Chanpaisit, Yaohui Wang, Xinyuan Chen, Limin Wang, Dahua Lin, Yu Qiao, and Ziwei Liu. VBench: Comprehensive benchmark suite for video generative mod- els....

    2024

  31. [38]

    NTIRE 2025 challenge on video quality enhancement for video conferencing: Datasets, methods and results

    Varun Jain, Zongwei Wu, Quan Zou, Louis Florentin, Henrik Turbell, Sandeep Siddhartha, Radu Timofte, et al. NTIRE 2025 challenge on video quality enhancement for video conferencing: Datasets, methods and results. InPro- ceedings of the IEEE/CVF Conference on Computer Vision an...

    2025

  32. [39]

    Vqa 2: Visual question an- swering for video quality assessment

    Ziheng Jia, Zicheng Zhang, Jiaying Qian, Haoning Wu, Wei Sun, Chunyi Li, Xiaohong Liu, Weisi Lin, Guang- tao Zhai, and Xiongkuo Min. Vqa 2: Visual question an- swering for video quality assessment. InarXiv preprint arXiv:2503.10078, 2025. 3

    2025 arXiv

  33. [40]

    Stablevqa: A deep no-reference quality assessment model for video stability

    Tengchuan Kou, Xiaohong Liu, Jun Jia, Wei Sun, Guangtao Zhai, and Ning Liu. Stablevqa: A deep no-reference quality assessment model for video stability. InProceedings of the 31st ACM International Conference on Multimedia, 2023. 2

    2023

  34. [41]

    Stablevqa: A deep no- reference quality assessment model for video stability

    Tengchuan Kou, Xiaohong Liu, Wei Sun, Jun Jia, Xiongkuo Min, Guangtao Zhai, and Ning Liu. Stablevqa: A deep no- reference quality assessment model for video stability. In Proceedings of the 31st ACM International Conference on Multimedia, pages 1066–1076, 2023. 2

    2023

  35. [43]

    Pika: Ai video generation platform

    Pika Labs. Pika: Ai video generation platform. In https://pika.art/, 2024. 4

    2024

  36. [44]

    NTIRE 2025 challenge on efficient burst hdr and restoration: Datasets, methods, and results

    Sangmin Lee, Eunpil Park, Angel Canelo, Hyunhee Park, Youngjo Kim, Hyungju Chun, Xin Jin, Chongyi Li, Chun- Le Guo, Radu Timofte, et al. NTIRE 2025 challenge on efficient burst hdr and restoration: Datasets, methods, and results. InProceedings of the IEEE/CVF Conference on Com...

    2025

  37. [46]

    NTIRE 2025 challenge on day and night raindrop removal for dual-focused images: Methods and results

    Xin Li, Yeying Jin, Xin Jin, Zongwei Wu, Bingchen Li, Yufei Wang, Wenhan Yang, Yu Li, Zhibo Chen, Bihan Wen, Robby Tan, Radu Timofte, et al. NTIRE 2025 challenge on day and night raindrop removal for dual-focused images: Methods and results. InProceedings of the IEEE/CVF Confe...

    2025

  38. [47]

    NTIRE 2025 challenge on short-form ugc video quality assessment and enhancement: Kwaisr dataset and study

    Xin Li, Xijun Wang, Bingchen Li, Kun Yuan, Yizhen Shao, Suhang Yao, Ming Sun, Chao Zhou, Radu Timofte, and Zhibo Chen. NTIRE 2025 challenge on short-form ugc video quality assessment and enhancement: Kwaisr dataset and study. InProceedings of the IEEE/CVF Conference on Compute...

    2025

  39. [48]

    NTIRE 2025 challenge on short-form ugc video qual- ity assessment and enhancement: Methods and results

    Xin Li, Kun Yuan, Bingchen Li, Fengbin Guan, Yizhen Shao, Zihao Yu, Xijun Wang, Yiting Lu, Wei Luo, Suhang Yao, Ming Sun, Chao Zhou, Zhibo Chen, Radu Timofte, et al. NTIRE 2025 challenge on short-form ugc video qual- ity assessment and enhancement: Methods and results. In Proc...

    2025

  40. [49]

    Ugc-video: Perceptual quality as- sessment of user-generated videos

    Yang Li, Shengbin Meng, Xinfeng Zhang, Shiqi Wang, Yue Wang, and Siwei Ma. Ugc-video: Perceptual quality as- sessment of user-generated videos. In2020 IEEE Confer- ence on Multimedia Information Processing and Retrieval (MIPR), pages 35–38. IEEE, 2020. 2

    2020

  41. [50]

    NTIRE 2025 the 2nd restore any image model (RAIM) in the wild challenge

    Jie Liang, Radu Timofte, Qiaosi Yi, Zhengqiang Zhang, Shuaizheng Liu, Lingchen Sun, Rongyuan Wu, Xindong Zhang, Hui Zeng, Lei Zhang, et al. NTIRE 2025 the 2nd restore any image model (RAIM) in the wild challenge. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion a...

    2025

  42. [51]

    NTIRE 2025 XGC quality assessment challenge: Methods and results

    Xiaohong Liu, Xiongkuo Min, Qiang Hu, Xiaoyun Zhang, Jie Guo, et al. NTIRE 2025 XGC quality assessment challenge: Methods and results. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2025. 2

    2025

  43. [52]

    Ntire 2023 quality assessment of video enhancement challenge

    Xiaohong Liu, Radu Timofte, Yunlong Dong, Zhiliang Ma, Haotian Fan, Chunzheng Zhu, Xiongkuo Min, Guangtao Zhai, Ziheng Jia, Mirko Agarla, et al. Ntire 2023 quality assessment of video enhancement challenge. InProceed- ings of the IEEE/CVF Conference on Computer Vision and Patt...

    2023

  44. [53]

    NTIRE 2025 challenge on low light image enhancement: Methods and results

    Xiaoning Liu, Zongwei Wu, Florin-Alexandru Vasluianu, Hailong Yan, Bin Ren, Yulun Zhang, Shuhang Gu, Le Zhang, Ce Zhu, Radu Timofte, et al. NTIRE 2025 challenge on low light image enhancement: Methods and results. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion ...

    2025

  45. [54]

    Evalcrafter: Benchmark- ing and evaluating large video generation models.arXiv preprint arXiv:2310.11440, 2023

    Yaofang Liu, Xiaodong Cun, Xuebo Liu, Xintao Wang, Yong Zhang, Haoxin Chen, Yang Liu, Tieyong Zeng, Ray- mond Chan, and Ying Shan. Evalcrafter: Benchmark- ing and evaluating large video generation models.arXiv preprint arXiv:2310.11440, 2023. 2, 3

    arXiv

  46. [55]

    Fetv: A bench- mark for fine-grained evaluation of open-domain text-to- video generation.Advances in Neural Information Process- ing Systems, 36, 2024

    Yuanxin Liu, Lei Li, Shuhuai Ren, Rundong Gao, Shicheng Li, Sishuo Chen, Xu Sun, and Lu Hou. Fetv: A bench- mark for fine-grained evaluation of open-domain text-to- video generation.Advances in Neural Information Process- ing Systems, 36, 2024. 2, 3

    2024

  47. [56]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision, pages 10012–10022, 2021. 8

    2021

  48. [57]

    Kvq: Kwai video quality assessment for short-form videos

    Yiting Lu, Xin Li, Yajing Pei, Kun Yuan, Qizhi Xie, Yun- peng Qu, Ming Sun, Chao Zhou, and Zhibo Chen. Kvq: Kwai video quality assessment for short-form videos. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, 2024. 2

    2024

  49. [58]

    Latte: La- tent diffusion transformer for video generation

    Xin Ma, Yaohui Wang, Gengyun Jia, Xinyuan Chen, Ziwei Liu, Yuanfang Li, Cunjian Chen, and Yu Qiao. Latte: La- tent diffusion transformer for video generation. InarXiv preprint:2401.03048, 2024. 4

    2024 arXiv

  50. [59]

    Dreamtalk: When expressive talking head generation meets diffusion probabilistic models.arXiv preprint arXiv:2312.09767, 2023

    Yifeng Ma et al. Dreamtalk: When expressive talking head generation meets diffusion probabilistic models.arXiv preprint arXiv:2312.09767, 2023. 3

    2023 arXiv

  51. [60]

    Video quality assessment on mobile devices: Subjective, behavioral and objective studies.IEEE Journal of Selected Topics in Signal Pro- cessing, 6(6):652–671, 2012

    Anush Krishna Moorthy, Lark Kwon Choi, Alan Conrad Bovik, and Gustavo De Veciana. Video quality assessment on mobile devices: Subjective, behavioral and objective studies.IEEE Journal of Selected Topics in Signal Pro- cessing, 6(6):652–671, 2012. 2

    2012

  52. [61]

    Human lips synchronisation in au- todesk maya

    Antonio Moura, Ingrida Mazonaviciute, Jo ˜ao Nunes, and Justinas Grigaravicius. Human lips synchronisation in au- todesk maya. In2007 14th International Workshop on Systems, Signals and Image Processing and 6th EURASIP Conference focused on Speech and Image Processing, Mul- ti...

    2007

  53. [62]

    A lip sync expert is all you need for speech to lip generation in the wild

    KR Prajwal, Rudrabha Mukhopadhyay, Vinay P Nambood- iri, and CV Jawahar. A lip sync expert is all you need for speech to lip generation in the wild. InACM MM, 2020. 3

    2020

  54. [63]

    The tenth NTIRE 2025 efficient super-resolution challenge report

    Bin Ren, Hang Guo, Lei Sun, Zongwei Wu, Radu Tim- ofte, Yawei Li, et al. The tenth NTIRE 2025 efficient super-resolution challenge report. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2025. 2

    2025

  55. [64]

    Pirenderer: Controllable portrait image generation via semantic neural rendering

    Yurui Ren, Ge Li, Yuanqi Chen, Thomas H Li, and Shan Liu. Pirenderer: Controllable portrait image generation via semantic neural rendering. InProceedings of the IEEE/CVF international conference on computer vision, pages 13759– 13768, 2021. 3

    2021

  56. [65]

    NTIRE 2025 challenge on UGC video enhancement: Meth- ods and results

    Nickolay Safonov, Alexey Bryntsev, Andrey Moskalenko, Dmitry Kulikov, Dmitriy Vatolin, Radu Timofte, et al. NTIRE 2025 challenge on UGC video enhancement: Meth- ods and results. InProceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR) Worksh...

    2025

  57. [66]

    NTIRE 2025 challenge on event-based image deblurring: Methods and results

    Lei Sun, Andrea Alfarano, Peiqi Duan, Shaolin Su, Kaiwei Wang, Boxin Shi, Radu Timofte, Danda Pani Paudel, Luc Van Gool, et al. NTIRE 2025 challenge on event-based image deblurring: Methods and results. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Re...

    2025

  58. [67]

    The tenth ntire 2025 image denoising challenge report

    Lei Sun, Hang Guo, Bin Ren, Luc Van Gool, Radu Timo- fte, Yawei Li, et al. The tenth ntire 2025 image denoising challenge report. InProceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR) Workshops, 2025. 2

    2025

  59. [68]

    A deep learning based no-reference quality assessment model for ugc videos

    Wei Sun, Xiongkuo Min, Wei Lu, and Guangtao Zhai. A deep learning based no-reference quality assessment model for ugc videos. InProceedings of the 30th ACM Interna- tional Conference on Multimedia, pages 856–865, 2022. 2

    2022

  60. [69]

    A deep learning based no-reference quality assessment model for ugc videos

    Wei Sun, Xiongkuo Min, Wei Lu, and Guangtao Zhai. A deep learning based no-reference quality assessment model for ugc videos. InProceedings of the 30th ACM Interna- tional Conference on Multimedia, page 856–865, 2022. 5, 7

    2022

  61. [70]

    Kling ai

    Kuaishou Team. Kling ai. Inhttps://klingai.io/, 2024. 4

    2024

  62. [71]

    Facegan: Facial attribute controllable reenactment gan

    Soumya Tripathy, Juho Kannala, and Esa Rahtu. Facegan: Facial attribute controllable reenactment gan. InProceed- ings of the IEEE/CVF winter conference on applications of computer vision, pages 1329–1338, 2021. 3

    2021

  63. [72]

    NTIRE 2025 image shadow removal challenge report

    Florin-Alexandru Vasluianu, Tim Seizinger, Zhuyun Zhou, Cailian Chen, Zongwei Wu, Radu Timofte, et al. NTIRE 2025 image shadow removal challenge report. InProceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2025. 2

    2025

  64. [73]

    NTIRE 2025 ambi- ent lighting normalization challenge

    Florin-Alexandru Vasluianu, Tim Seizinger, Zhuyun Zhou, Zongwei Wu, Radu Timofte, et al. NTIRE 2025 ambi- ent lighting normalization challenge. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2025. 2

    2025

  65. [74]

    MCL-JCV: a jnd-based h

    Haiqiang Wang, Weihao Gan, Sudeng Hu, Joe Yuchieh Lin, Lina Jin, Longguang Song, Ping Wang, Ioannis Kat- savounidis, Anne Aaron, and C-C Jay Kuo. MCL-JCV: a jnd-based h. 264/avc video quality assessment dataset. In Proceedings of the IEEE International Conference on Im- age Pr...

    2016

  66. [75]

    Eca-net: Efficient channel attention for deep convolutional neural networks

    Qilong Wang, Banggu Wu, Pengfei Zhu, Peihua Li, Wang- meng Zuo, and Qinghua Hu. Eca-net: Efficient channel attention for deep convolutional neural networks. In2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 11531–11539, 2020. 8

    2020

  67. [76]

    Safa: Structure aware face animation

    Qiulin Wang, Lu Zhang, and Bo Li. Safa: Structure aware face animation. In2021 International Conference on 3D Vision (3DV), pages 679–688. IEEE, 2021. 3

    2021

  68. [77]

    Audio2head: Audio-driven one-shot talking- head generation with natural head motion.arXiv preprint arXiv:2107.09293, 2021

    Suzhen Wang, Lincheng Li, Yu Ding, Changjie Fan, and Xin Yu. Audio2head: Audio-driven one-shot talking- head generation with natural head motion.arXiv preprint arXiv:2107.09293, 2021. 3

    2021 arXiv

  69. [78]

    Hififace: 3d shape and semantic prior guided high fidelity face swapping.arXiv preprint arXiv:2106.09965, 2021

    Yuhan Wang, Xu Chen, Junwei Zhu, Wenqing Chu, Ying Tai, Chengjie Wang, Jilin Li, Yongjian Wu, Feiyue Huang, and Rongrong Ji. Hififace: 3d shape and semantic prior guided high fidelity face swapping.arXiv preprint arXiv:2106.09965, 2021. 3

    2021 arXiv

  70. [79]

    Rich features for perceptual quality assessment of ugc videos

    Yilin Wang, Junjie Ke, Hossein Talebi, Joong Gon Yim, Neil Birkbeck, Balu Adsumilli, Peyman Milanfar, and Feng Yang. Rich features for perceptual quality assessment of ugc videos. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 13435–...

    2021

  71. [80]

    NTIRE 2025 challenge on light field image super-resolution: Methods and results

    Yingqian Wang, Zhengyu Liang, Fengyuan Zhang, Lvli Tian, Longguang Wang, Juncheng Li, Jungang Yang, Radu Timofte, Yulan Guo, et al. NTIRE 2025 challenge on light field image super-resolution: Methods and results. InPro- ceedings of the IEEE/CVF Conference on Computer Vision an...

    2025

  72. [81]

    Image quality assessment: from error visibility to structural similarity.IEEE transactions on image pro- cessing, 13(4):600–612, 2004

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. Image quality assessment: from error visibility to structural similarity.IEEE transactions on image pro- cessing, 13(4):600–612, 2004. 3

    2004

  73. [82]

    X2face: A network for controlling face generation using images, au- dio, and pose codes

    Olivia Wiles, A Koepke, and Andrew Zisserman. X2face: A network for controlling face generation using images, au- dio, and pose codes. InProceedings of the European con- ference on computer vision (ECCV), pages 670–686, 2018. 3

    2018

  74. [83]

    Fast-vqa: Efficient end-to-end video quality assessment with fragment sampling

    Haoning Wu, Chaofeng Chen, Jingwen Hou, Liang Liao, Annan Wang, Wenxiu Sun, Qiong Yan, and Weisi Lin. Fast-vqa: Efficient end-to-end video quality assessment with fragment sampling. InProceedings of the Computer Vision–ECCV 2022: 17th European Conference, Tel Aviv, Israel, Oct...

    2022

  75. [84]

    Fast- vqa: Efficient end-to-end video quality assessment with fragment sampling, 2022

    Haoning Wu, Chaofeng Chen, Jingwen Hou, Liang Liao, Annan Wang, Wenxiu Sun, Qiong Yan, and Weisi Lin. Fast- vqa: Efficient end-to-end video quality assessment with fragment sampling, 2022. 5, 7

    2022

  76. [85]

    Towards explainable in-the-wild video quality assess- ment: a database and a language-prompted approach

    Haoning Wu, Erli Zhang, Liang Liao, Chaofeng Chen, Jing- wen Hou, Annan Wang, Wenxiu Sun, Qiong Yan, and Weisi Lin. Towards explainable in-the-wild video quality assess- ment: a database and a language-prompted approach. In Proceedings of the 31st ACM International Conference ...

    2023

  77. [86]

    Exploring video quality assessment on user generated contents from aesthetic and technical per- spectives

    Haoning Wu, Erli Zhang, Liang Liao, Chaofeng Chen, Jing- wen Hou Hou, Annan Wang, Wenxiu Sun Sun, Qiong Yan, and Weisi Lin. Exploring video quality assessment on user generated contents from aesthetic and technical per- spectives. InInternational Conference on Computer Vision ...

    2023

  78. [87]

    Q-align: Teaching lmms for visual scoring via discrete text-defined levels.arXiv preprint arXiv:2312.17090, 2023

    Haoning Wu, Zicheng Zhang, Weixia Zhang, Chaofeng Chen, Chunyi Li, Liang Liao, Annan Wang, Erli Zhang, Wenxiu Sun, Qiong Yan, Xiongkuo Min, Guangtao Zhai, and Weisi Lin. Q-align: Teaching lmms for visual scoring via discrete text-defined levels.arXiv preprint arXiv:2312.17090,...

    2023 arXiv

  79. [88]

    Towards open-ended visual quality comparison

    Haoning Wu, Hanwei Zhu, Zicheng Zhang, Erli Zhang, Chaofeng Chen, Liang Liao, Chunyi Li, et al. Towards open-ended visual quality comparison. InarXiv preprint arXiv:2403.11956, 2024. 3

    2024 arXiv

  80. [89]

    Reenactgan: Learning to reenact faces via boundary transfer

    Wayne Wu, Yunxuan Zhang, Cheng Li, Chen Qian, and Chen Change Loy. Reenactgan: Learning to reenact faces via boundary transfer. InProceedings of the European con- ference on computer vision (ECCV), pages 603–619, 2018. 3

    2018

  81. [90]

    NTIRE 2025 challenge on single image reflection removal in the wild: Datasets, methods and results

    Kangning Yang, Jie Cai, Ling Ouyang, Florin-Alexandru Vasluianu, Radu Timofte, Jiaming Ding, Huiming Sun, Lan Fu, Jinlong Li, Chiu Man Ho, Zibo Meng, et al. NTIRE 2025 challenge on single image reflection removal in the wild: Datasets, methods and results. InProceedings of the...

    2025

  82. [91]

    Cogvideox: Text-to- video diffusion models with an expert transformer

    Zhuoyi Yang, Jiayan Teng, Wendi Zheng, Ming Ding, Shiyu Huang, Jiazheng Xu, Yuanming Yang, Wenyi Hong, Xiaohan Zhang, and Guanyu Feng. Cogvideox: Text-to- video diffusion models with an expert transformer. InarXiv preprint:2408.06072, 2024. 4

    2024 arXiv

  83. [92]

    Mesh guided one-shot face reenactment using graph convo- lutional networks

    Guangming Yao, Yi Yuan, Tianjia Shao, and Kun Zhou. Mesh guided one-shot face reenactment using graph convo- lutional networks. InProceedings of the 28th ACM interna- tional conference on multimedia, pages 1773–1781, 2020. 3

    2020

  84. [93]

    Using brain- computer interface technology for modeling 3d objects in blender software.Journal of Automation Mobile Robotics and Intelligent Systems, 14, 2020

    Mateusz Zajac and Szczepan Paszkiel. Using brain- computer interface technology for modeling 3d objects in blender software.Journal of Automation Mobile Robotics and Intelligent Systems, 14, 2020. 3

    2020

  85. [94]

    Few-shot adversarial learning of real- istic neural talking head models

    Egor Zakharov, Aliaksandra Shysheya, Egor Burkov, and Victor Lempitsky. Few-shot adversarial learning of real- istic neural talking head models. InProceedings of the IEEE/CVF international conference on computer vision, pages 9459–9468, 2019. 3

    2019

  86. [95]

    NTIRE 2025 challenge on hr depth from images of specular and transparent surfaces

    Pierluigi Zama Ramirez, Fabio Tosi, Luigi Di Stefano, Radu Timofte, Alex Costanzino, Matteo Poggi, Samuele Salti, Stefano Mattoccia, et al. NTIRE 2025 challenge on hr depth from images of specular and transparent surfaces. InProceedings of the IEEE/CVF Conference on Computer V...

    2025

  87. [96]

    Sadtalker: Learning realistic 3d mo- tion coefficients for stylized audio-driven single image talk- ing face animation

    Wenxuan Zhang et al. Sadtalker: Learning realistic 3d mo- tion coefficients for stylized audio-driven single image talk- ing face animation. InIEEE/CVF CVPR, 2023. 3

    2023

  88. [97]

    Blind image quality assessment via vision- language correspondence: A multitask learning perspec- tive

    Weixia Zhang, Guangtao Zhai, Ying Wei, Xiaokang Yang, and Kede Ma. Blind image quality assessment via vision- language correspondence: A multitask learning perspec- tive. InIEEE Conference on Computer Vision and Pattern Recognition, pages 14071–14081, 2023. 7

    2023

  89. [98]

    Paps-ovqa: Projection-aware patch sampling for omnidirectional video quality assessment

    Zicheng Zhang, Yu Fan, Wei Sun, Xiongkuo Min, Xiao- hong Liu, Chunyi Li, Haoning Wu, Weisi Lin, Ning Liu, and Guangtao Zhai. Paps-ovqa: Projection-aware patch sampling for omnidirectional video quality assessment. In IEEE International Symposium on Circuits and Systems,

  90. [99]

    Dinet: Deformation inpainting network for realistic face visually dubbing on high resolu- tion video.arXiv preprint arXiv:2303.03988, 2023

    Zhimeng Zhang, Zhipeng Hu, Wenjin Deng, Changjie Fan, Tangjie Lv, and Yu Ding. Dinet: Deformation inpainting network for realistic face visually dubbing on high resolu- tion video.arXiv preprint arXiv:2303.03988, 2023. 3

    2023 arXiv

  91. [100]

    Q-bench-video: Benchmarking the video quality understanding of lmms

    Zicheng Zhang, Ziheng Jia, Haoning Wu, Chunyi Li, Zijian Chen, Yingjie Zhou, Wei Sun, Xiaohong Liu, Xiongkuo Min, Weisi Lin, and Guangtao Zhai. Q-bench-video: Benchmarking the video quality understanding of lmms. In arXiv preprint arXiv:2409.20063, 2024. 3

    2024 arXiv

  92. [101]

    Q-eval-100k: Evaluating visual quality and align- ment level for text-to-vision content

    Zicheng Zhang, Tengchuan Kou, Shushi Wang, Chunyi Li, Wei Sun, Wei Wang, Xiaoyu Li, Zongyu Wang, Xuezhi Cao, Xiongkuo Min, Xiaohong Liu, and Guangtao Zhai. Q-eval-100k: Evaluating visual quality and align- ment level for text-to-vision content. InarXiv preprint arXiv:2503.0235...

    2025 arXiv

  93. [102]

    A perceptual qual- ity assessment exploration for aigc images

    Zicheng Zhang, Chunyi Li, Wei Sun, Xiaohong Liu, Xiongkuo Min, and Guangtao Zhai. A perceptual qual- ity assessment exploration for aigc images. In2023 IEEE International Conference on Multimedia and Expo Work- shops (ICMEW), pages 440–445. IEEE, 2023. 3

    2023

  94. [103]

    Advancing zero-shot digital human qual- ity assessment through text-prompted evaluation.arXiv preprint arXiv:2307.02808, 2023

    Zicheng Zhang, Wei Sun, Yingjie Zhou, Haoning Wu, Chunyi Li, Xiongkuo Min, Xiaohong Liu, Guangtao Zhai, and Weisi Lin. Advancing zero-shot digital human qual- ity assessment through text-prompted evaluation.arXiv preprint arXiv:2307.02808, 2023. 2, 3

    2023 arXiv

  95. [104]

    Q-boost: On vi- sual quality assessment ability of low-level multi-modality foundation models

    Zicheng Zhang, Haoning Wu, Zhongpeng Ji, Chunyi Li, Erli Zhang, Wei Sun, Xiaohong Liu, et al. Q-boost: On vi- sual quality assessment ability of low-level multi-modality foundation models. InarXiv preprint arXiv:2312.15300,

    arXiv

  96. [105]

    Md-vqa: Multi-dimensional quality assessment for ugc live videos

    Zicheng Zhang, Wei Wu, Wei Sun, Danyang Tu, Wei Lu, Xiongkuo Min, Ying Chen, and Guangtao Zhai. Md-vqa: Multi-dimensional quality assessment for ugc live videos. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 1746–1755,

  97. [106]

    A reduced-reference quality assessment metric for textured mesh digital humans

    Zicheng Zhang, Yingjie Zhou, Chunyi Li, Kang Fu, Wei Sun, Xiaohong Liu, Xiongkuo Min, and Guangtao Zhai. A reduced-reference quality assessment metric for textured mesh digital humans. InICASSP 2024-2024 IEEE Inter- national Conference on Acoustics, Speech and Signal Pro- cess...

    2024

  98. [107]

    Quality assessment in the era of large mod- els: A survey

    Zicheng Zhang, Yingjie Zhou, Chunyi Li, Baixuan Zhao, Yixuan Gao, Zicheng Zhang, Chunyi Li, Haoning Wu, and Guangtao Zhai. Quality assessment in the era of large mod- els: A survey. InarXiv preprint arXiv:2409.00031, 2024. 3

    2024 arXiv

  99. [108]

    Ddh-qa: A dynamic digital humans quality assessment database

    Zicheng Zhang, Yingjie Zhou, Wei Sun, Wei Lu, Xiongkuo Min, Yu Wang, and Guangtao Zhai. Ddh-qa: A dynamic digital humans quality assessment database. In2023 IEEE International Conference on Multimedia and Expo (ICME), pages 2519–2524. IEEE, 2023. 3

    2023

  100. [109]

    Geometry-aware video quality as- sessment for dynamic digital human

    Zicheng Zhang, Yingjie Zhou, Wei Sun, Xiongkuo Min, and Guangtao Zhai. Geometry-aware video quality as- sessment for dynamic digital human. In2023 IEEE In- ternational Conference on Image Processing (ICIP), pages 1365–1369. IEEE, 2023. 3

    2023

  101. [110]

    Quality-of-experience evaluation for digital twins in 6g net- work environments.IEEE Transactions on Broadcasting, 70(3):995–1007, 2024

    Zicheng Zhang, Yingjie Zhou, Long Teng, Wei Sun, Chunyi Li, Xiongkuo Min, Xiao-Ping Zhang, and Guangtao Zhai. Quality-of-experience evaluation for digital twins in 6g net- work environments.IEEE Transactions on Broadcasting, 70(3):995–1007, 2024. 3

    2024

  102. [111]

    Identity- preserving talking face generation with landmark and ap- pearance priors

    Weizhi Zhong, Chaowei Fang, Yinqi Cai, Pengxu Wei, Gangming Zhao, Liang Lin, and Guanbin Li. Identity- preserving talking face generation with landmark and ap- pearance priors. InIEEE/CVF CVPR, 2023. 3

    2023

  103. [112]

    Light-vqa+: A video quality assess- ment model for exposure correction with vision-language guidance

    Xunchu Zhou, Xiaohong Liu, Yunlong Dong, Tengchuan Kou, Yixuan Gao, Zicheng Zhang, Chunyi Li, Haoning Wu, and Guangtao Zhai. Light-vqa+: A video quality assess- ment model for exposure correction with vision-language guidance. InarXiv preprint arXiv:2405.03333, 2024. 3

    2024 arXiv

  104. [113]

    An implementation of multimodal fusion sys- tem for intelligent digital human generation.arXiv preprint arXiv:2310.20251, 2023

    Yingjie Zhou, Yaodong Chen, Kaiyue Bi, Lian Xiong, and Hui Liu. An implementation of multimodal fusion sys- tem for intelligent digital human generation.arXiv preprint arXiv:2310.20251, 2023. 3

    2023 arXiv

  105. [114]

    Pyramid feature aggregation for hi- erarchical quality prediction of stitched panoramic images

    Yu Zhou, Weikang Gong, Yanjing Sun, Leida Li, Jinjian Wu, and Xinbo Gao. Pyramid feature aggregation for hi- erarchical quality prediction of stitched panoramic images. IEEE Transactions on Multimedia, 25:4177–4186, 2023. 8

    2023

  106. [115]

    Makelttalk: speaker-aware talking-head animation.ACM TOG, 2020

    Yang Zhou, Xintong Han, Eli Shechtman, Jose Echevar- ria, Evangelos Kalogerakis, and Dingzeyu Li. Makelttalk: speaker-aware talking-head animation.ACM TOG, 2020. 3

    2020

  107. [116]

    Memo-bench: A multiple benchmark for text-to-image and multimodal large language models on hu- man emotion analysis.arXiv preprint arXiv:2411.11235,

    Yingjie Zhou, Zicheng Zhang, Jiezhang Cao, Jun Jia, Yan- wei Jiang, Farong Wen, Xiaohong Liu, Xiongkuo Min, and Guangtao Zhai. Memo-bench: A multiple benchmark for text-to-image and multimodal large language models on hu- man emotion analysis.arXiv preprint arXiv:2411.11235,

    arXiv

  108. [117]

    Thqa: A perceptual quality assessment database for talking heads

    Yingjie Zhou, Zicheng Zhang, Wei Sun, Xiaohong Liu, Xiongkuo Min, Zhihua Wang, Xiao-Ping Zhang, and Guangtao Zhai. Thqa: A perceptual quality assessment database for talking heads. In2024 IEEE International Conference on Image Processing (ICIP), pages 15–21. IEEE, 2024. 2, 3, 4

    2024

  109. [118]

    Subjective and objec- tive quality-of-experience assessment for 3d talking heads

    Yingjie Zhou, Zicheng Zhang, Wei Sun, Xiaohong Liu, Xiongkuo Min, and Guangtao Zhai. Subjective and objec- tive quality-of-experience assessment for 3d talking heads. InProceedings of the 32nd ACM International Conference on Multimedia, pages 6033–6042, 2024. 2, 3, 4

    2024

  110. [119]

    A no-reference quality as- sessment method for digital human head

    Yingjie Zhou, Zicheng Zhang, Wei Sun, Xiongkuo Min, Xi- anghe Ma, and Guangtao Zhai. A no-reference quality as- sessment method for digital human head. In2023 IEEE In- ternational Conference on Image Processing (ICIP), pages 36–40. IEEE, 2023. 3

    2023

  111. [120]

    3dgcqa: A quality assessment database for 3d ai- generated contents

    Yingjie Zhou, Zicheng Zhang, Farong Wen, Jun Jia, Yan- wei Jiang, Xiaohong Liu, Xiongkuo Min, and Guangtao Zhai. 3dgcqa: A quality assessment database for 3d ai- generated contents. InICASSP 2025-2025 IEEE Interna- tional Conference on Acoustics, Speech and Signal Pro- cessing...

    2025

  112. [121]

    Reli-qa: A multidimensional quality assessment dataset for relighted human heads

    Yingjie Zhou, Zicheng Zhang, Farong Wen, Jun Jia, Xiongkuo Min, Jia Wang, and Guangtao Zhai. Reli-qa: A multidimensional quality assessment dataset for relighted human heads. In2024 IEEE International Conference on Visual Communications and Image Processing (VCIP), pages 1–5. ...

    2024

  113. [122]

    Esvqa: Percep- tual quality assessment of egocentric spatial videos.arXiv preprint arXiv:2412.20423, 2024

    Xilei Zhu, Huiyu Duan, Liu Yang, Yucheng Zhu, Xiongkuo Min, Guangtao Zhai, and Patrick Le Callet. Esvqa: Percep- tual quality assessment of egocentric spatial videos.arXiv preprint arXiv:2412.20423, 2024. 2

    2024 arXiv

  114. [123]

    Controllable radiance fields for dynamic face synthesis

    Peiye Zhuang, Liqian Ma, Sanmi Koyejo, and Alexander Schwing. Controllable radiance fields for dynamic face synthesis. In2022 International Conference on 3D Vision (3DV), pages 1–11. IEEE, 2022. 3

    2022

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.