# T-FAKE: Synthesizing Thermal Images for Facial Landmarking Source: [T-FAKE: Synthesizing Thermal Images for Facial Landmarking]() Collection index: [llms.txt]() Data updated: 2026-01-04T05:02:40.756Z Authors: Philipp Flotho, Moritz Piening, Anna Kukleva, Gabriele Steidl (2) \_\_INSTITUTION\_4\_\_ Saarland University, (2) Institute of Mathematics, Technische Universit\\"at Berlin, (3) Max Planck Institute for Informatics, Saarland Informatics Campus) Published: 2024-08-27T15:07:58.000Z ## Abstract Facial analysis is a key component in a wide range of applications such as healthcare, autonomous driving, and entertainment. Despite the availability of various facial RGB datasets, the thermal modality, which plays a crucial role in life sciences, medicine, and biometrics, has been largely overlooked. To address this gap, we introduce the T-FAKE dataset, a new large-scale synthetic thermal dataset with sparse and dense landmarks. To facilitate the creation of the dataset, we propose a novel RGB2Thermal loss function, which enables the domain-adaptive transfer of RGB faces to thermal style. By utilizing the Wasserstein distance between thermal and RGB patches and the statistical analysis of clinical temperature distributions on faces, we ensure that the generated thermal images closely resemble real samples. Using RGB2Thermal style transfer based on our RGB2Thermal loss function, we create the large-scale synthetic thermal T-FAKE dataset with landmark and segmentation annotations. Leveraging our novel T-FAKE dataset, probabilistic landmark prediction, and label adaptation networks, we demonstrate significant improvements in landmark detection methods on thermal images across different landmark conventions. Our models show excellent performance with both sparse 70-point landmarks and dense 478-point landmark annotations. Moreover, our RGB2Thermal loss leads to notable results in terms of perceptual evaluation and temperature prediction. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Facial Landmark Detection on CHARLOTTE Low 1.0](): Facial Landmark Detection; CHARLOTTE Low 1.0; NME (W/H): 0.1018. - [Facial Landmark Detection on CHARLOTTE 1.0 (Side)](): Facial Landmark Detection; CHARLOTTE 1.0 (Side); NME W/H: 0.0551. - [Facial Landmark Detection on CHARLOTTE 1.0 (Front)](): Facial Landmark Detection; CHARLOTTE 1.0 (Front); NME W/H: 0.1059. - [Facial Landmark Detection on CHARLOTTE 1.0 (Full)](): Facial Landmark Detection; CHARLOTTE 1.0 (Full); NME W/H: 0.0815. - [Facial Landmark Detection on CHARLOTTE High 1.0](): Facial Landmark Detection; CHARLOTTE High 1.0; NME (W/H): 0.0661. - [Facial Landmarking on 300W Full set (test)](): Facial Landmarking; 300W Full set (test); NME IO: 0.0522. - [Facial Landmarking on 300W Challenging (test)](): Facial Landmarking; 300W Challenging (test); NME IO: 0.0743. - [Facial Landmark Detection on 300-W (Common)](): Facial Landmark Detection; 300-W (Common); NME: not recorded. ## Original sources and implementation [Original paper]() [PDF]() [Implementation]()