ArcFace: Additive Angular Margin Loss for Deep Face Recognition
Citations
29 citations
Cites methods from "ArcFace: Additive Angular Margin Lo..."
...[64] used a perceptual loss that minimised the cosine distance of image features extracted from Iin and Imod using the ArcFace network [71]....
[...]
...between id-features extracted from ArcFace [71]....
[...]
29 citations
29 citations
29 citations
Cites background or methods from "ArcFace: Additive Angular Margin Lo..."
...BASELINE LONG-TERM PEDESTRIAN RE-IDENTIFICATION PERFORMANCE OBTAINED BY AN ENSEMBLE OF ARCFACE [8] + COSAM [31] IN THE P-DESTRE DATA SET...
[...]
...7), from where a feature representation was obtained using the ArcFace [8] model....
[...]
...The inner plot provides the top-20 results as a zoomed-in region. during the test phase, the mean value of the ArcFace facial features in the tracklet were appended to the body-based representation yielding from COSAM....
[...]
...The facial regions-of-interest were detected by the SSH method [25] (acceptance threshold=0.7), from where a feature representation was obtained using the ArcFace [8] model....
[...]
...As for the previous tasks, the full specification of the samples used in each split is given in.15 For the ArcFace method, the MobileNetV2 was used as backbone model, and the learning rate set to 0.01....
[...]
29 citations
Cites methods from "ArcFace: Additive Angular Margin Lo..."
...We establish baselines of face recognition on several advanced and effective architectures [9, 32, 11, 23], as prior models, based on the criterion of ArcFace [5] with m = 0.5....
[...]
...We establish baselines of face recognition on several advanced and effective architectures [9, 32, 11, 23], as prior models, based on the criterion of ArcFace [5] with m = 0....
[...]
References
123,388 citations
33,597 citations
30,843 citations
13,268 citations
10,447 citations