Gaussian Vector: An Efficient Solution for Facial Landmark Detection

Xiong, Yilin; Zhou, Zijian; Dou, Yuhao; Su, Zhizhong

doi:10.1007/978-3-030-69541-5_5

Cited by 12 publications

(16 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this section, we compare the proposed LOTR with several state-of-the-art methods, including Look-at-Boundary (LAB) [38], Wing loss [39], adaptive Wing loss (AWing) [54], LUVLi [47], Gaussian vector (GV) [34], and Heatmap-In-Heatmap (HIH) [53]. As shown in Table 2, our proposed LOTR-HR+ achieves an NME of 4.31%, clearly outperforming LAB, Wing, AWing, and LUVLi methods, and yields an AUC of 60.14%, surpassing all state-of-the-arts by a large margin (0.44-6.91 points).…”

Section: Results On the Wflw Datasetmentioning

confidence: 99%

“…• We demonstrate that the proposed LOTR framework detects facial landmarks accurately. Experimental results indicate the superiority of the proposed LOTR over other algorithms on the leaderboard of the First JD-landmark localization challenge and two recent heatmap-based methods [1], [34]. On another benchmark, WFLW dataset [38], the results show the proposed LOTR method is comparable with several stateof-the-art methods.…”

mentioning

confidence: 76%

“…Xiong et al [34] proposed a vectorization approach in which each set of vector labels represents a ground-truth landmark's position using a quasi-Gaussian distribution. This proposed distribution is a Gaussian density function enhancing the distribution peak with an additional constant θ while being bounded by a threshold 3σ.…”

Section: B Heatmap Regressionmentioning

confidence: 99%

“…Following [1], [34], [47], [54], we used the standard metrics including the normalized mean error (NME), the failure rate, and the area under the curve (AUC) of the cumulative distribution to evaluate and compare landmark localization algorithms. The NME is computed across all predicted landmarks as follows:…”

Section: B Evaluation Metricsmentioning

confidence: 99%

See 3 more Smart Citations

LOTR: Face Landmark Localization Using Localization Transformer

Watchareeruetai¹,

Sommanna²,

Jain³

et al. 2022

IEEE Access

View full text Add to dashboard Cite

Section: Results On the Wflw Datasetmentioning

confidence: 99%

mentioning

confidence: 76%

Section: B Heatmap Regressionmentioning

confidence: 99%

Section: B Evaluation Metricsmentioning

confidence: 99%

See 2 more Smart Citations

LOTR: Face Landmark Localization Using Localization Transformer

Watchareeruetai¹,

Sommanna²,

Jain³

et al. 2022

IEEE Access

View full text Add to dashboard Cite

“…Another viable way to obtain efficient and robust model is designing specific modules or architectures. Xiong et al [12] proposed the Gaussian vector to reduce the model complexity. Lee et al [13] exploited the advantage of geometric prior-generative adversarial network to design an associated learning framework for facial landmark detection.…”

Section: Introductionmentioning

confidence: 99%

Robust and Efficient Facial Landmark Localization

Hu¹,

Wang

Tao³

et al. 2021

2021 IEEE International Conference on Multimedia &Amp; Expo Workshops (ICMEW)

View full text Add to dashboard Cite

Facial landmark localization is an essential component for facial recognition which is a vital technology for patients tracking in the situation of a global pandemic of COVID-19. Parameter amount and computation cost are two key factors affecting the application of facial landmark localization. In addition, the intrinsic imbalance of head pose in existing datasets has a great impact on the model generalization. To address these problems, this paper proposed a lightweight model LiteFace with multi-knowledge distillation and a pose-aware resampling strategy. Specifically, we proposed to extract multi-scale features and enhance the receptive filed of network by delicate model design to obtain an efficient architecture and improved the performance of the lightweight model by multi-knowledge distillation which combines knowledge distillation and mutual learning in a two-stage manner. To further improve the model performance and boost the generalization, we proposed a pose-aware resampling strategy that generates samples with different head poses and utilized the LaPa dataset to generate masked face images to increase data diversity. We conducted extensive ablation studies on the model design, multi-knowledge distillation and resampling strategy. The proposed method achieves 1.93% and 3.04% NME on the JD-landmark-mask val and test dataset respectively. Finally, LiteFace wins the third place in the 3rd Grand Challenge of 106-Point Facial Landmark Localization.

show abstract

SimCC: A Simple Coordinate Classification Perspective for Human Pose Estimation

Yang

Liu

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Gaussian Vector: An Efficient Solution for Facial Landmark Detection

Cited by 12 publications

References 45 publications

LOTR: Face Landmark Localization Using Localization Transformer

LOTR: Face Landmark Localization Using Localization Transformer

Robust and Efficient Facial Landmark Localization

SimCC: A Simple Coordinate Classification Perspective for Human Pose Estimation

Contact Info

Product

Resources

About