Pith. sign in

REVIEW 2 major objections 6 minor 60 references

Towards Unified Dynamic Face Landmark Detection

T0 review · 2 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read A single model can be trained on any mixture of face landmark datasets and answer arbitrary landmark queries at runtime.

desk verdict A solid incremental contribution to FLD unification; the central dynamic-query evidence is currently undermined by an ambiguous held-out protocol that needs clarification. read the letter →

arxiv 2608.10346 v1 pith:SE7ODJ45 submitted 2026-08-11 cs.CV cs.AI

Sebastian Regalado , Varshanth R. Rao , Ruowei Jiang , Parham Aarabi , Igor Gilitschenski This is my paper · ORCID
classification cs.CVcs.AI
keywords facelandmarkdetectionPart-AnchoredPositionsunifiedtemplatedynamicpredictioncross-datasetgeneralizationtransformerdecoderqueriesalignment
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

Face landmark datasets disagree on how many points define a face and what each point means, so today one model must be trained per dataset and it can only emit that dataset's fixed set of points. This paper tries to remove both restrictions with a single representation: every landmark is described as a progression value between 0 and 1 along the contour of a face part (eye, lip, jawline, etc.), a value the authors call FPALP. Because any landmark from any dataset can be written this way, the datasets can be merged and one network trained on the union; at runtime, the user supplies FPALP-based queries for exactly the landmarks wanted, and the same network predicts them. The paper shows this unified dynamic model matches or beats prior specialist models on WFLW, 300W, and AFLW-19 while also predicting landmarks that were never seen during training, such as denser or partial face-part layouts. A reader should care because this collapses many per-dataset models into one trainable, queryable detector and points toward a common interface for face geometry.

What carries the argument

The load-bearing object is the Face Part-Anchored Landmark Position (FPALP), a normalized coordinate that locates any facial landmark as a value in [0,1] representing its fractional distance from the start to the end of a user-defined face-part contour (eyes, brows, lips, nose, face contour). To build it across datasets, the authors first make a unified face template T_U by clustering the landmarks of AFLW-19, 300W, and WFLW, yielding tight clusters with average intra-cluster distance 2.22 pixels; each cluster becomes a position on a face-part curve. Closed contours duplicate their start landmark as the end so that fraction 0 and 1 meet. That FPALP value, added to a text embedding of the face part name, forms an image-agnostic query; a cross-modality decoder, combining self-attention, deformable image cross-attention, and text cross-attention, refines the query and its coordinate prediction over several layers. The FPALP is what lets different annotation schemes share one parameter set, and the text-plus-progression query is what lets the number of output landmarks be chosen freely at runtime.

What would settle it

Take a fourth landmark dataset whose contour ordering or point spacing differs from the three used here, align it into the unified template, and measure the average intra-cluster distance; if it climbs well above the reported 2.22 pixels, or if a model trained on that fusion loses its edge over per-dataset models on cross-template evaluations, the semantic-alignment premise fails.

Watch

Extended reading notes

Core claim

The central claim is that a single end-to-end model can be trained on the fusion of multiple N-point face landmark datasets and then, at inference, answer arbitrary landmark queries, producing any number of landmarks the user asks for. The discovery supporting this is that the landmark definitions of AFLW-19, 300W, and WFLW, though formally different, are semantically aligned through shared face parts: the same physical points recur across datasets and are roughly evenly spaced along each part's contour. The paper encodes each landmark as an FPALP, a fraction between 0 and 1 of the way along its containing face-part curve, and uses the FPALP plus the face part's text embedding as a query. A decoder refines the query against the image and outputs coordinates. On the full benchmarks the model matches or outperforms prior per-dataset specialists (e.g., 4.05 NME on WFLW without dataset adapters, 2.47 on 300W common, 1.02 NMEdiag on AFLW-19), and with lightweight per-dataset adapters it improves further, while a 300W-only model transfers to unseen templates such as WFLW68 and even to the 28 WFLW landmarks absent from 300W.

Load-bearing premise

The method's success rests on the assumption that aligning the landmark templates of different datasets into one shared template keeps each landmark's meaning intact; if landmarks that should be the same point land far apart in the merged template, the unified training signal becomes contradictory.

Editorial extensions

If this is right

  • One set of weights trains jointly on any number of N-point landmark datasets, so adding a dataset no longer means training a new backbone or regression head.
  • At inference the same network can emit sparse or dense layouts, from four points per face part to high granularity multipliers, by loading different FPALP queries, without retraining.
  • Joint training on varied templates regularizes the model: fused-dataset training generally beats single-dataset training and transfers to unseen templates in near zero-shot fashion.
  • With small per-dataset adapters, the unified model's numbers surpass the reported specialist SOTA on WFLW, 300W, and AFLW-19, showing that the shared representation retains dataset-specific precision.
  • Because each query is a face-part phrase plus a progression value, landmark detection becomes addressable by language-like instructions, the interface the paper says it aims to enable.

Reading between the lines

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

  • If the FPALP alignment stays semantically consistent as more datasets are added, the unified template could keep absorbing new annotation schemes without retraining from scratch, a scale-up the paper only demonstrates for three datasets.
  • The same query mechanism points toward open-vocabulary face parts: a stronger multilingual text encoder might let a user name a novel facial region and receive landmarks for it, which the paper lists as future work.
  • A sharp stress test would be to fuse a dataset whose contour ordering runs opposite to the three used here; if the average intra-cluster distance grows well beyond 2.22 pixels, unified training should lose its edge over per-dataset models.
  • Because the paper left COFW out of training after observing degradation, the method's tolerance for annotation noise is not yet characterized; auditing label quality before fusion may be essential in practice.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 6 minor

Summary. The paper proposes Face Part-Anchored Landmark Positions (FPALPs), in which each facial landmark is represented as a normalized progression value along a user-defined face-part contour, together with a query-based transformer decoder that takes FPALP-plus-text embeddings as input and predicts the corresponding landmark coordinates. The authors claim two system-level benefits: unified training, in which one model can be trained end-to-end on a fusion of differently annotated face-landmark datasets (AFLW19, 300W, WFLW), and dynamic prediction, in which landmarks can be queried at arbitrary FPALPs at inference time without retraining. Experiments compare the method with state-of-the-art detectors on WFLW, 300W, AFLW-19, and cross-dataset targets (COFW, COFW68, WFLW68), ablate the training data, text encoder, image encoder, and decoder depth, and include a held-out landmark experiment against cubic-spline interpolation.

Significance. If the claims hold, this is a practically useful step toward a single face-landmark model that is not bound to one N-point annotation protocol: it removes the need for per-dataset heads, supports query-driven landmark selection, and makes no use of 3D priors. The paper is unusually transparent about its limitations, including template-alignment imprecision, the inability to evaluate interpolated density benchmarks, and language bias in the text encoder. The cross-dataset zero-shot evaluation on WFLWE (Table 6b) is a valuable falsifiable design, and the dataset-ablation study in Table 4 directly probes the unified-training hypothesis. However, the head-to-head evidence against geometric interpolation in Table 7 is currently ambiguous because the masking protocol may leak FPALPs across datasets, and all performance claims rest on single-run point estimates without released code or checkpoints.

major comments (2)
  1. [§A.6, Table 7; Tables 12–13] The controlled held-out landmark experiment does not state whether the reduced-supervision model is trained on the fused AFLW19+300W+WFLW set or on a single dataset. Under fused training, the withheld FPALPs are not necessarily unseen: Tables 12 and 13 show that every 300W face-contour FPALP (0/32, 2/32, …, 32/32) and all 300W outer/inner-lip FPALPs also occur in WFLW, and most 300W eyebrow, nose-bridge, and nose-boundary FPALPs coincide as well. A model trained on the fusion could therefore memorize the exact FPALP-to-coordinate mapping from another dataset, and the reported 19.0%, 16.3%, and 15.7% improvements over cubic splines would be inflated. The authors should explicitly state the training protocol for the reduced-supervision rows, and if fused training was used, re-run the experiment with a global FPALP blacklist across all datasets, or use single-dataset training for this comparison.
  2. [Table 2; Section 4.1] All quantitative results are single-run point estimates without standard deviations, confidence intervals, or significance tests, and the SOTA comparison in Table 2 cites previously reported numbers without shared evaluation infrastructure or code release. Since several claimed advantages are small (for example, 2.80 vs. 2.78 NME on 300W, and 4.05 vs. 4.02 on WFLW), the paper should report multiple seeds, provide error bars, or explicitly temper the claim of consistent superiority over prior methods; otherwise the comparative claims are not fully supported.
minor comments (6)
  1. [§A.16] The sentence before Tables 10–13 says the mapping is tabulated for “the AFLW, COFW, 300W, and 300W datasets,” but the fourth dataset should be WFLW, and “Tab. Table 10-13” is ungrammatical.
  2. [Table 1] The column headers in Table 1 run together (“TrainingCycles InferenceCalculation StorageParameterLandmarkThroughput”); they need spacing to be readable.
  3. [§A.9, Eq. (8)] Equation (8) should define the terms “RelativePosition” and “index” formally and state explicitly that the denominator Np in FPALP = pos/(Np−1) is the number of landmarks in the unified face-part template Tp, since the current notation is ambiguous about whether Np refers to the dataset-specific or unified template.
  4. [§A.6, Table 7] The description “alternating landmark indices within the face-part definitions” should specify whether the alternation is applied to the original dataset ID ordering or to the sorted FPALP ordering; these two orderings differ for non-monotone contours such as closed eye and lip curves.
  5. [Table 6b and §A.6] The set of 28 WFLW landmarks defining WFLWE should be enumerated or otherwise unambiguously specified, because some WFLW positions share an FPALP value with a 300W position (for example, the eye point at 4/8 equals the 300W eye point at 3/6), and it is not clear whether those shared-value points are included in the “absent/undefined” set.
  6. [General] The paper would benefit from releasing the FPALP mapping tables, dataset-adapter configurations, and trained checkpoints; given that the method depends on the manual face-part and start/end ordering choices, code or detailed mapping files are important for reproducing Table 2 and Table 7.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: FPALPs are a re-parameterization of supplied annotations, and all central claims are evaluated against external benchmarks.

full rationale

The paper's central derivation chain is self-contained: FPALPs are defined directly from each dataset's native landmark indices via FPALP_l,p = pos_l,p / (Np - 1), so the representation is a normalization of the input annotations rather than a fitted quantity that is later renamed as a prediction. The coordinate predictions are produced by a learned image-conditioned decoder supervised with Wing loss and PossLoss, and the unified/dynamic capabilities are architectural properties of the query construction, not consequences of an equation whose output is its own input. All principal performance claims are measured against external benchmarks (WFLW, 300W, AFLW-19, COFW, COFW68, WFLW68), with results compared to previously reported SOTA numbers; no benchmark result is derived from the FPALP definition itself. The controlled held-out experiment in Section A.6/Table 7, while it could have been clearer about whether the reduced-supervision rows use single-dataset or fused training, does not reduce to construction: the reported full-supervision NMEs (3.56 for 300W, 4.51 and 5.32 for WFLW) are markedly worse than the fused-training results in the main paper (2.80 and 4.05), indicating those rows are not the fused model, so the 19.0%, 16.3%, and 15.7% gains over cubic-spline interpolation are not forced by the training protocol as written. The paper's self-citations appear only in non-load-bearing related-work contexts and no uniqueness theorem or ansatz is imported from the authors' prior work to justify the central design. The paper also honestly acknowledges its own limitations, including approximate template alignment and the inability to evaluate on interpolation-derived dense benchmarks, which further supports that the claims are not circular.

Assumptions & free parameters 2 free parameters · 4 assumptions · 1 invented entities

The paper's contribution rests on representational assumptions about how landmarks relate to face part boundaries, plus the architectural hypothesis that query-conditioned decoding works. All are empirically motivated within the paper: the alignment statistics, text encoder ablation, and held-out dynamic prediction experiment. No physical free constants are fitted. Hyperparameters are standard except ndec, which is ablated.

free parameters (2)
  • ndec (decoder layers) = 3
    Ablated in Table 8 and fixed for all reported results; peripheral to the central claim.
  • Dataset Adapter LoRA rank = 4
    Used only for the adapter variant that claims state-of-the-art surpass; not part of the unified model.
assumptions (4)
  • domain assumption Landmarks in the considered N-point datasets are semantically anchored to face parts and approximately evenly spaced along face part boundaries.
    Underpins the FPALP definition; introduced in Sec. 3 and Figure 2, citing LAB [39] and LDDMM-Face [42].
  • domain assumption The union of face templates from AFLW19, 300W, and WFLW clusters into tight, semantically consistent face-part templates with mean intra-cluster distance 2.22 pixels, and approximate alignment is sufficient.
    Load-bearing for unified training; described in Sec. 3 and A.15.
  • domain assumption Pretrained text encoder embeddings of face part phrases carry semantics useful for landmark localization.
    Supported by the ablation in Table 5; used in Eq. 1.
  • domain assumption A query decoder can map FPALP plus text queries to coordinates and generalize to unseen FPALP values.
    Core model hypothesis; tested via held-out landmark experiments in A.6.
invented entities (1)
  • Face Part-Anchored Landmark Positions (FPALPs)
    purpose: To encode any facial landmark as a normalized progression value along a user-defined face part contour, enabling unified and dynamic queries.
    A representational device introduced by the paper, not an empirical entity. Its usefulness is measured through the paper's internal experiments; no external falsifiable handle independent of the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Towards Unified Dynamic Face Landmark Detection." pith.science (2026). https://pith.science/paper/SE7ODJ45

@misc{pith2026260810346,
  author       = {Pith},
  title        = {Pith review of: Towards Unified Dynamic Face Landmark Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SE7ODJ45}},
  note         = {Machine review of arXiv:2608.10346}
}
abstract

Although advancements in face landmark detection (FLD) methods continue to push performance boundaries, they overlook two major functional limitations: (1) different network parameters need to be trained independently for each ``$N$-point'' benchmark dataset, and (2) a model trained on an ``$N$-point'' dataset reliably outputs only the $N$ landmarks. In our work, we first conceptualize Face Part-Anchored Landmark Positions (FPALPs), wherein each landmark is treated as a progression value between zero (start) and one (end) along a face part's contour. Every landmark can be expressed in the FPALP format, irrespective of its source dataset, hence unlocking the ability to unify all ``$N$-point'' datasets into a single dataset. Secondly, we represent each landmark with an FPALP-based query, refine it progressively with a cross-modality decoder, and predict its coordinates based on the final representation. Our approach, called Unified Dynamic FLD, embodies these two design choices and streamlines the landmark detection pipeline by enabling (1) a single model to learn on any number of ``$N$-point'' datasets, and (2) yield any number of specific landmark predictions by loading the designated landmark queries at runtime. Extensive experiments on multiple benchmark datasets show that our method delivers these benefits while remaining competitive with, and in several cases outperforming existing state-of-the-art methods.

Figures

Figures reproduced from arXiv: 2608.10346 by the authors.

Figure 1
Figure 1. A comparison of the end-to-end train￾ing pipeline of prior works’ separate model and common backbone paradigms to the sin￾gle model paradigm implemented by our Uni￾fied Dynamic Face Landmark Detection method. BB, RH, and D denote backbone, regression head, and number of datasets, respectively. Based on Face Part-Anchored Landmark Posi￾tions, our network can train on the combination of multiple “N-point" datasets and… view at source ↗
Figure 3
Figure 3. An overview of our proposed framework. First, we associate user-defined face parts to [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Illustration of the construction of Face [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figures from the paper (5 more)
Figure 5
Figure 5. Figure 5: A comparison of training curve plots when using learnable embeddings vs. Sentence￾BERT [30] to represent face parts [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: An illustration of the dynamic landmark prediction capability of our system. Images are [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: An illustration of the dynamic landmark prediction capability of our system. All images [PITH_FULL_IMAGE:figures/full_fig_p020_7.png]
Figure 8
Figure 8. Figure 8: An illustration of the dynamic landmark prediction capability of our system on challenging [PITH_FULL_IMAGE:figures/full_fig_p021_8.png]
Figure 9
Figure 9. Figure 9: An illustration of the dynamic landmark prediction capability of our system on challenging [PITH_FULL_IMAGE:figures/full_fig_p022_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

60 extracted references · 56 canonical work pages

  1. [1]

    Al-Nuimi and Ghassan J

    Arqam M. Al-Nuimi and Ghassan J. Mohammed. Face direction estimation based on mediapipe landmarks. InICCITM, 2021

    work page 2021

  2. [2]

    Bougourzi, F

    F. Bougourzi, F. Dornaika, and A. Taleb-Ahmed. Deep learning based face beauty prediction via dynamic robust losses and ensemble regression.Knowledge-Based Systems, 2022

    work page 2022

  3. [3]

    Burgos-Artizzu, Pietro Perona, and Piotr Dollár

    Xavier P. Burgos-Artizzu, Pietro Perona, and Piotr Dollár. Robust face landmark estimation under occlusion. InICCV, 2013

    work page 2013

  4. [4]

    3d face reconstruction and dense alignment with a new generated dataset.Displays, 2021

    Mingcheng Cai, Shuo Zhang, Guoqiang Xiao, and Shoucheng Fan. 3d face reconstruction and dense alignment with a new generated dataset.Displays, 2021

    work page 2021

  5. [5]

    Continuous landmark detection with 3d queries

    Prashanth Chandran, Gaspard Zoss, Paulo Gotardo, and Derek Bradley. Continuous landmark detection with 3d queries. InCVPR, 2023

    work page 2023

  6. [6]

    An image is worth 16x16 words: Transformers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition at scale. InICLR, 2021

    2021

  7. [7]

    Wing loss for robust facial landmark localisation with convolutional neural networks

    Zhen-Hua Feng, Josef Kittler, Muhammad Awais, Patrik Huber, and Xiao-Jun Wu. Wing loss for robust facial landmark localisation with convolutional neural networks. InCVPR, 2018

    work page 2018

  8. [8]

    Zhang, Shaoqing Ren, and Jian Sun

    Kaiming He, X. Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. InCVPR, 2016

    work page 2016

Show all 60 references
  1. [9]

    LoRA: Low-rank adaptation of large language models

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. LoRA: Low-rank adaptation of large language models. InICLR, 2022

    2022

  2. [10]

    Adnet: Leveraging error-bias towards normal direction in face alignment

    Yangyu Huang, Hao Yang, Chong Li, Jongyoo Kim, and Fangyun Wei. Adnet: Leveraging error-bias towards normal direction in face alignment. InICCV, 2021

    2021

  3. [11]

    Freeenricher: Enriching face landmarks without additional cost.AAAI, 2023

    Yangyu Huang, Xi Chen, Jongyoo Kim, Hao Yang, Chong Li, Jiaolong Yang, and Dong Chen. Freeenricher: Enriching face landmarks without additional cost.AAAI, 2023

    2023

  4. [12]

    Iyer, Rahul K., Ruban Nersisson, Zhemin Zhuang, Alex Noel Joseph Raj, and Imthiaz Refayee

    Tharun J. Iyer, Rahul K., Ruban Nersisson, Zhemin Zhuang, Alex Noel Joseph Raj, and Imthiaz Refayee. Machine learning-based facial beauty prediction and analysis of frontal facial images using facial landmarks and traditional image descriptors.Computational Intelligence and Ne...

    2021

  5. [13]

    Pixel-in-pixel net: Towards efficient facial landmark detection in the wild.IJCV, 2021

    Haibo Jin, Shengcai Liao, and Ling Shao. Pixel-in-pixel net: Towards efficient facial landmark detection in the wild.IJCV, 2021

    2021

  6. [14]

    Towards stabilizing facial landmark detection and tracking via hierarchical filtering: A new method.Journal of the Franklin Institute, 2020

    Yi Jin, Xingyan Guo, Yidong Li, Junliang Xing, and Hui Tian. Towards stabilizing facial landmark detection and tracking via hierarchical filtering: A new method.Journal of the Franklin Institute, 2020. ISSN 0016-0032

    2020

  7. [15]

    Pintavirooj

    Aniwat Juhong and C. Pintavirooj. Face recognition based on facial landmark detection. In BMEiCON, 2017

    2017

  8. [16]

    Deep graphics encoder for real-time video makeup synthesis from example

    Robin Kips, Ruowei Jiang, Sileye Ba, Edmund Phung, Parham Aarabi, Pietro Gori, Matthieu Perrot, and Isabelle Bloch. Deep graphics encoder for real-time video makeup synthesis from example. InCVPRW, 2021

    2021

  9. [17]

    Marks, Wenxuan Mou, Ye Wang, Michael Jones, Anoop Cherian, Toshiaki Koike-Akino, Xiaoming Liu, and Chen Feng

    Abhinav Kumar, Tim K. Marks, Wenxuan Mou, Ye Wang, Michael Jones, Anoop Cherian, Toshiaki Koike-Akino, Xiaoming Liu, and Chen Feng. Luvli face alignment: Estimating landmarks’ location, uncertainty, and visibility likelihood. InCVPR, 2020

    2020

  10. [18]

    Towards accurate facial landmark detection via cascaded transformers

    Hui Li, Zidong Guo, Seon-Min Rhee, Seungju Han, and Jae-Joon Han. Towards accurate facial landmark detection via cascaded transformers. InCVPR, June 2022. 10

    2022

  11. [19]

    Lightweight real-time makeup try-on in mobile browsers with tiny cnn models for facial tracking

    TianXing Li, Zhi Yu, Edmund Phung, Brendan Duke, Irina Kezele, and Parham Aarabi. Lightweight real-time makeup try-on in mobile browsers with tiny cnn models for facial tracking. InCVPRW, 2019

    2019

  12. [20]

    Structured landmark detection via topology-adapting deep graph learning

    Weijian Li, Yuhang Lu, Kang Zheng, Haofu Liao, Chihung Lin, Jiebo Luo, Chi-Tung Cheng, Jing Xiao, Le Lu, Chang-Fu Kuo, et al. Structured landmark detection via topology-adapting deep graph learning. InECCV, 2020

    2020

  13. [21]

    Grounding dino: Marrying dino with grounded pre-training for open-set object detection

    Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chun yue Li, Jianwei Yang, Hang Su, Jun-Juan Zhu, and Lei Zhang. Grounding dino: Marrying dino with grounded pre-training for open-set object detection. InECCV, 2023

    2023

  14. [22]

    Designing an AI-based virtual try-on web application.Sensors, 2022

    Davide Marelli, Simone Bianco, and Gianluigi Ciocca. Designing an AI-based virtual try-on web application.Sensors, 2022

    2022

  15. [23]

    M. I. N. P. Munasinghe. Facial expression recognition using facial landmarks and random forest classifier. InICIS, 2018

    2018

  16. [24]

    Facexformer: A unified transformer for facial analysis.arXiv preprint arXiv:2403.12960, 2024

    Kartik Narayan, Vibashan VS, Rama Chellappa, and Vishal M Patel. Facexformer: A unified transformer for facial analysis.arXiv preprint arXiv:2403.12960, 2024

    2024 arXiv

  17. [25]

    Facial landmark-based emotion recognition via directed graph neural network.Electronics, 2020

    Quang Tran Ngoc, Seunghyun Lee, and Byung Cheol Song. Facial landmark-based emotion recognition via directed graph neural network.Electronics, 2020

    2020

  18. [26]

    Aggregation via separation: Boosting facial landmark detector with semi-supervised style translation

    Shengju Qian, Keqiang Sun, Wayne Wu, Chen Qian, and Jiaya Jia. Aggregation via separation: Boosting facial landmark detector with semi-supervised style translation. InICCV, 2019

    2019

  19. [27]

    Faceptor: A generalist model for face perception

    Lixiong Qin, Mei Wang, Xuannan Liu, Yuhang Zhang, Wei Deng, Xiaoshuai Song, Weiran Xu, and Weihong Deng. Faceptor: A generalist model for face perception. InECCV, 2024

    2024

  20. [28]

    Hyperface: A deep multi-task learning framework for face detection, landmark localization, pose estimation, and gender recognition

    Rajeev Ranjan, Vishal Patel, and Rama Chellappa. Hyperface: A deep multi-task learning framework for face detection, landmark localization, pose estimation, and gender recognition. TPAMI, 2016

    2016

  21. [29]

    Castillo, and Rama Chellappa

    Rajeev Ranjan, Swami Sankaranarayanan, Carlos D. Castillo, and Rama Chellappa. An all-in- one convolutional neural network for face analysis. InFG, 2017

    2017

  22. [30]

    Sentence-bert: Sentence embeddings using siamese bert- networks

    Nils Reimers and Iryna Gurevych. Sentence-bert: Sentence embeddings using siamese bert- networks. InEMNLP, 2019

    2019

  23. [31]

    300 faces in-the-wild challenge: The first facial landmark localization challenge

    Christos Sagonas, Georgios Tzimiropoulos, Stefanos Zafeiriou, and Maja Pantic. 300 faces in-the-wild challenge: The first facial landmark localization challenge. InICCVW, June 2013

    2013

  24. [32]

    Angel Arul Jothi

    Divija Sanapala and J. Angel Arul Jothi. Virtual makeup try-on system using cognitive learning. InCognitive Analytics and Reinforcement Learning, 2024. ISBN 9781394214068

    2024

  25. [33]

    Face recognition based on facial landmarks

    Adil Sarsenov and Konstantin Latuta. Face recognition based on facial landmarks. InAICT, 2017

    2017

  26. [34]

    Nicuface: Robust neonatal face detection in complex nicu scenes.IEEE Access, 2022

    Yasmina Souley Dosso, Daniel Kyrollos, Kim Greenwood, Joann Harrold, and James Green. Nicuface: Robust neonatal face detection in complex nicu scenes.IEEE Access, 2022

    2022

  27. [35]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Ł ukasz Kaiser, and Illia Polosukhin. Attention is all you need. InNeurIPS, 2017

    2017

  28. [36]

    Interpretable facial landmark detection by multi-expert collaborative uncertainty-aware deep networks.Neural Networks, 2026

    Jun Wan, Hui Xi, Yuanzhi Yao, Hang Sun, Zhihui Lai, and Jie Zhou. Interpretable facial landmark detection by multi-expert collaborative uncertainty-aware deep networks.Neural Networks, 2026

    2026

  29. [37]

    3d face reconstruction with dense landmarks

    Erroll Wood, Tadas Baltrušaitis, Charlie Hewitt, Matthew Johnson, Jingjing Shen, Nikola Milosavljevi´c, Daniel Wilde, Stephan Garbin, Toby Sharp, Ivan Stojiljkovi´c, Tom Cashman, and Julien Valentin. 3d face reconstruction with dense landmarks. InECCV, 2022

    2022

  30. [38]

    Design of a facial landmark detection system using a dynamic optical flow approach.IEEE Access, 2021

    Bing-Fei Wu, Bo-Rui Chen, and Chun-Fei Hsu. Design of a facial landmark detection system using a dynamic optical flow approach.IEEE Access, 2021. 11

    2021

  31. [39]

    Look at boundary: A boundary-aware face alignment algorithm

    Wayne Wu, Chen Qian, Shuo Yang, Quan Wang, Yici Cai, and Qiang Zhou. Look at boundary: A boundary-aware face alignment algorithm. InCVPR, 2018

    2018

  32. [40]

    Sparse local patch transformer for robust face alignment and landmarks inherent relation learning

    Jiahao Xia, Weiwei Qu, Wenjian Huang, Jianguo Zhang, Xi Wang, and Min Xu. Sparse local patch transformer for robust face alignment and landmarks inherent relation learning. InCVPR, 2022

    2022

  33. [41]

    Anchorface: An anchor-based facial landmark detector across large poses

    Zixuan Xu, Banghuai Li, Miao Geng, Ye Yuan, and Gang Yu. Anchorface: An anchor-based facial landmark detector across large poses. InAAAI, 2020

    2020

  34. [42]

    Lddmm-face: Large deformation diffeomorphic metric learning for cross-annotation face alignment.Pattern Recog- nition, 2024

    Huilin Yang, Junyan Lyu, Pujin Cheng, Roger Tam, and Xiaoying Tang. Lddmm-face: Large deformation diffeomorphic metric learning for cross-annotation face alignment.Pattern Recog- nition, 2024. ISSN 0031-3203

    2024

  35. [43]

    Wider face: A face detection benchmark

    Shuo Yang, Ping Luo, Chen Change Loy, and Xiaoou Tang. Wider face: A face detection benchmark. InCVPR, 2016

    2016

  36. [44]

    Sparse to dense motion transfer for face image animation

    Ruiqi Zhao, Tianyi Wu, and Guodong Guo. Sparse to dense motion transfer for face image animation. InICCVW, 2021

    2021

  37. [45]

    General facial representation learning in a visual- linguistic manner

    Yinglin Zheng, Hao Yang, Ting Zhang, Jianmin Bao, Dongdong Chen, Yangyu Huang, Lu Yuan, Dong Chen, Ming Zeng, and Fang Wen. General facial representation learning in a visual- linguistic manner. InCVPR, June 2022

    2022

  38. [46]

    STAR Loss: Reducing semantic ambiguity in facial landmark detection

    Zhenglin Zhou, Huaxia Li, Hong Liu, Nanyang Wang, Gang Yu, and Rongrong Ji. STAR Loss: Reducing semantic ambiguity in facial landmark detection. InCVPR, 2023

    2023

  39. [47]

    Possloss: A reliable and sensitive facial landmark detection loss function

    Qikui Zhu. Possloss: A reliable and sensitive facial landmark detection loss function. InICCV, 2025

    2025

  40. [48]

    Face alignment by coarse-to-fine shape searching

    Shizhan Zhu, Cheng Li, Chen Change Loy, and Xiaoou Tang. Face alignment by coarse-to-fine shape searching. InCVPR, 2015

    2015

  41. [49]

    smiling girl with party wig

    Xizhou Zhu, Weijie Su, Lewei Lu, Bin Li, Xiaogang Wang, and Jifeng Dai. Deformable DETR: deformable transformers for end-to-end object detection. InICLR, 2021. 12 A Technical Appendices and Supplementary Material A.1 Relevance and Strength of Contribution Our work addresses cr...

    2021

  42. [50]

    This provides a fully supervised reference

    Full supervision:the model is trained using the complete 68- or 98-point landmark template and evaluated only on the designated held-out subset. This provides a fully supervised reference

  43. [51]

    Reduced supervision:the model is trained using only the retained landmarks and is directly queried at the FPALPs corresponding to the held-out landmarks

  44. [52]

    Predict 10 coordinates from the left chin boundary to the end of the jawline,

    Cubic-spline interpolation:cubic splines are fitted independently for each face part using the retained landmark predictions from the reduced-supervision model and are sampled at the FPALPs corresponding to the held-out landmarks. We use cubic rather than linear interpolation ...

  45. [53]

    Soft spatial attention maps based on the introduced face parts can be used to approximate the boundaries of unseen face parts, even under occlusion

    We can utilize the text encoder to parse face part descriptions into latent embeddings that can be aligned with image features. Soft spatial attention maps based on the introduced face parts can be used to approximate the boundaries of unseen face parts, even under occlusion. ...

  46. [54]

    This allows the framework to learn robust part representations even when portions of the face are not visible

    We can also leverage visibility annotations per landmark, such as those provided in the MERL-RA V [17] dataset, to supervise the model in learning to selectively ignore occluded regions during training. This allows the framework to learn robust part representations even when p...

  47. [55]

    This can lower computational costs and facilitate more accessible deployment of FLD systems, particularly in resource-constrained settings

    Accessibility and Efficiency: By unifying multiple landmark datasets into a single frame- work, the method reduces the need to train separate models for different datasets. This can lower computational costs and facilitate more accessible deployment of FLD systems, particularl...

  48. [56]

    Scalability and Adaptability: The ability to handle variable landmark layouts allows for more adaptable systems that can be tailored for specific applications, such as facial expression analysis (select face parts only), face direction estimation, FLD stabilization in videos, ...

  49. [57]

    Potential for Improved Fairness: Training a single model on diverse datasets may mitigate biases that arise from models trained solely on specific datasets, potentially leading to more robust performance across varied demographic and environmental conditions. 18

  50. [58]

    Negative Impacts:

    Cross-Dataset Learning: The FPALP framework can promote cross-dataset learning, en- couraging researchers to leverage underutilized datasets and discover new patterns in facial landmark configurations. Negative Impacts:

  51. [59]

    Dataset Bias Amplification: If the unified dataset disproportionately represents certain demographics or facial structures, the model could inadvertently reinforce existing biases, leading to inaccurate predictions or unfair outcomes

  52. [60]

    If certain datasets contain noisy or imprecise labels, the model’s predictions may propagate these errors, potentially compromising its generalizabil- ity and reliability

    Dependence on Dataset Quality: The proposed approach relies on the accuracy and con- sistency of landmark annotations. If certain datasets contain noisy or imprecise labels, the model’s predictions may propagate these errors, potentially compromising its generalizabil- ity and...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.