Papers › Learning an Animatable Detailed 3D Face Model from In-The-Wild Images

Learning an Animatable Detailed 3D Face Model from In-The-Wild Images

7 Dec 2020arXiv:2012.04012archive 2025-07-28

Yao Feng, Haiwen Feng, Michael J. Black, Timo Bolkart

While current monocular 3D face reconstruction methods can recover fine geometric details, they suffer several limitations. Some methods produce faces that cannot be realistically animated because they do not model how wrinkles vary with expression. Other methods are trained on high-quality face scans and do not generalize well to in-the-wild images. We present the first approach that regresses 3D face shape and animatable details that are specific to an individual but change with expression. Our model, DECA (Detailed Expression Capture and Animation), is trained to robustly produce a UV displacement map from a low-dimensional latent representation that consists of person-specific detail parameters and generic expression parameters, while a regressor is trained to predict detail, shape, albedo, expression, pose and illumination parameters from a single image. To enable this, we introduce a novel detail-consistency loss that disentangles person-specific details from expression-dependent wrinkles. This disentanglement allows us to synthesize realistic person-specific wrinkles by controlling expression parameters while keeping person-specific details unchanged. DECA is learned from in-the-wild images with no paired 3D supervision and achieves state-of-the-art shape reconstruction accuracy on two benchmarks. Qualitative results on in-the-wild data demonstrate DECA's robustness and its ability to disentangle identity- and expression-dependent details enabling animation of reconstructed faces. The model and code are publicly available at https://deca.is.tue.mpg.de.

PaperPDFCode

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

Code

YadiraF/DECA officialmentioned in papermentioned on GitHubpytorch report
yfeng95/deca mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

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

Tasks

3D Face Alignment3D Face Animation3D Face Modelling3D Face ReconstructionDisentanglementFace AlignmentFace ModelFace Reconstruction

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
3D Face Reconstruction NoW Benchmark DECA Mean Reconstruction Error (mm) 1.38 #5 of 17 Archive leaderboard report
3D Face Reconstruction NoW Benchmark DECA Median Reconstruction Error 1.09 #5 of 17 Archive leaderboard report
3D Face Reconstruction NoW Benchmark DECA Stdev Reconstruction Error (mm) 1.18 #5 of 17 Archive leaderboard report
3D Face Reconstruction REALY DECA-c @cheek 1.479 (±0.535) #14 of 24 Archive leaderboard report
3D Face Reconstruction REALY DECA-c @forehead 2.394 (±0.576) #14 of 24 Archive leaderboard report
3D Face Reconstruction REALY DECA-c @mouth 2.516 (±0.839) #14 of 24 Archive leaderboard report
3D Face Reconstruction REALY DECA-c @nose 1.697 (±0.355) #14 of 24 Archive leaderboard report
3D Face Reconstruction REALY DECA-c all 2.010 #14 of 24 Archive leaderboard report
3D Face Reconstruction REALY DECA-f @cheek 1.443 (±0.498) #20 of 24 Archive leaderboard report
3D Face Reconstruction REALY DECA-f @forehead 2.457 (±0.559) #20 of 24 Archive leaderboard report
3D Face Reconstruction REALY DECA-f @mouth 2.802 (±0.868) #20 of 24 Archive leaderboard report
3D Face Reconstruction REALY DECA-f @nose 2.138 (±0.461) #20 of 24 Archive leaderboard report
3D Face Reconstruction REALY DECA-f all 2.210 #20 of 24 Archive leaderboard report
3D Face Reconstruction REALY (side-view) DECA-c @cheek 1.630 (±1.135) #13 of 19 Archive leaderboard report
3D Face Reconstruction REALY (side-view) DECA-c @forehead 2.423 (±0.720) #13 of 19 Archive leaderboard report
3D Face Reconstruction REALY (side-view) DECA-c @mouth 2.472 (±1.079) #13 of 19 Archive leaderboard report
3D Face Reconstruction REALY (side-view) DECA-c @nose 1.903 (±1.050) #13 of 19 Archive leaderboard report
3D Face Reconstruction REALY (side-view) DECA-c all 2.107 #13 of 19 Archive leaderboard report
3D Face Reconstruction REALY (side-view) DECA-f @cheek 1.555 (±0.822) #17 of 19 Archive leaderboard report
3D Face Reconstruction REALY (side-view) DECA-f @forehead 2.519 (±0.718) #17 of 19 Archive leaderboard report
3D Face Reconstruction REALY (side-view) DECA-f @mouth 2.684 (±1.041) #17 of 19 Archive leaderboard report
3D Face Reconstruction REALY (side-view) DECA-f @nose 2.286 (±1.103) #17 of 19 Archive leaderboard report
3D Face Reconstruction REALY (side-view) DECA-f all 2.261 #17 of 19 Archive leaderboard report
3D Face Reconstruction Stirling-HQ (FG2018 3D face reconstruction challenge) DECA Mean Reconstruction Error (mm) 1.89 #1 of 4 Archive leaderboard report
3D Face Reconstruction Stirling-LQ (FG2018 3D face reconstruction challenge) DECA Mean Reconstruction Error (mm) 1.91 #1 of 4 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.

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