Papers › Subpixel Heatmap Regression for Facial Landmark Localization

Subpixel Heatmap Regression for Facial Landmark Localization

3 Nov 2021arXiv:2111.02360archive 2025-07-28

Adrian Bulat, Enrique Sanchez, Georgios Tzimiropoulos

Deep Learning models based on heatmap regression have revolutionized the task of facial landmark localization with existing models working robustly under large poses, non-uniform illumination and shadows, occlusions and self-occlusions, low resolution and blur. However, despite their wide adoption, heatmap regression approaches suffer from discretization-induced errors related to both the heatmap encoding and decoding process. In this work we show that these errors have a surprisingly large negative impact on facial alignment accuracy. To alleviate this problem, we propose a new approach for the heatmap encoding and decoding process by leveraging the underlying continuous distribution. To take full advantage of the newly proposed encoding-decoding mechanism, we also introduce a Siamese-based training that enforces heatmap consistency across various geometric image transformations. Our approach offers noticeable gains across multiple datasets setting a new state-of-the-art result in facial landmark localization. Code alongside the pretrained models will be made available at https://www.adrianbulat.com/face-alignment

PaperPDF

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

No code repository is listed for this paper in the archive or in Syntology's graph.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Face Alignmentregression

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Face Alignment 300W SHR-FAN NME_inter-ocular (%, Challenge) 4.13 #8 of 48 Archive leaderboard report
Face Alignment 300W SHR-FAN NME_inter-ocular (%, Common) 2.61 #8 of 48 Archive leaderboard report
Face Alignment 300W SHR-FAN NME_inter-ocular (%, Full) 2.94 #8 of 48 Archive leaderboard report
Face Alignment 300W Split 2 (300W-LP) SH-FAN AUC@7 (bbox) 71.1 #1 of 4 Archive leaderboard report
Face Alignment 300W Split 2 (300W-LP) SH-FAN NME (bbox) 2.04 #1 of 4 Archive leaderboard report
Face Alignment 300W Split 2 (300W-LP) SH-FAN NME (inter-ocular) 2.94 #1 of 4 Archive leaderboard report
Face Alignment AFLW-19 SHR-FAN AUC_box@0.07 (%, Full) 70.0 #5 of 23 Archive leaderboard report
Face Alignment AFLW-19 SHR-FAN NME_box (%, Full) 2.14 #5 of 23 Archive leaderboard report
Face Alignment AFLW-19 SHR-FAN NME_diag (%, Frontal) 1.12 #5 of 23 Archive leaderboard report
Face Alignment AFLW-19 SHR-FAN NME_diag (%, Full) 1.31 #5 of 23 Archive leaderboard report
Face Alignment COFW SH-FAN NME (inter-ocular) 3.02% #2 of 28 Archive leaderboard report
Face Alignment COFW-68 (300WLP) SH-FAN AUC@7 64.9 #1 of 4 Archive leaderboard report
Face Alignment COFW-68 (300WLP) SH-FAN NME (box) 2.47 #1 of 4 Archive leaderboard report
Face Alignment WFLW SH-FAN AUC@10 (inter-ocular) 63.81 #1 of 36 Archive leaderboard report
Face Alignment WFLW SH-FAN FR@10 (inter-ocular) 1.55 #1 of 36 Archive leaderboard report
Face Alignment WFLW SH-FAN NME (inter-ocular) 3.72 #1 of 36 Archive leaderboard report
Face Alignment WFW (Extra Data) SH-FAN AUC@10 (inter-ocular) 63.1 #1 of 11 Archive leaderboard report
Face Alignment WFW (Extra Data) SH-FAN FR@10 (inter-ocular) 1.55 #1 of 11 Archive leaderboard report
Face Alignment WFW (Extra Data) SH-FAN NME (inter-ocular) 3.72 #1 of 11 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Heatmap

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections