Papers › EdgeFace: Efficient Face Recognition Model for Edge Devices

EdgeFace: Efficient Face Recognition Model for Edge Devices

4 Jul 2023arXiv:2307.01838archive 2025-07-28

Anjith George, Christophe Ecabert, Hatef Otroshi Shahreza, Ketan Kotwal, Sebastien Marcel

In this paper, we present EdgeFace, a lightweight and efficient face recognition network inspired by the hybrid architecture of EdgeNeXt. By effectively combining the strengths of both CNN and Transformer models, and a low rank linear layer, EdgeFace achieves excellent face recognition performance optimized for edge devices. The proposed EdgeFace network not only maintains low computational costs and compact storage, but also achieves high face recognition accuracy, making it suitable for deployment on edge devices. Extensive experiments on challenging benchmark face datasets demonstrate the effectiveness and efficiency of EdgeFace in comparison to state-of-the-art lightweight models and deep face recognition models. Our EdgeFace model with 1.77M parameters achieves state of the art results on LFW (99.73%), IJB-B (92.67%), and IJB-C (94.85%), outperforming other efficient models with larger computational complexities. The code to replicate the experiments will be made available publicly.

PaperPDFCodeCode Syntology ran

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

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="2307.01838")

Code

Syntology Ran 2 of 2 code samples harvested from 1 repository linked to this paper; 0 have no recorded run. Of those that ran: 2 ran · our draft was wrong.

By repository: official repository: 2 samples from 1 repository, 2 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

otroshi/edgeface officialmentioned 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

2 samples harvested; 2 ran; 0 honoured the contract we drafted; 0 have no recorded run. Read from Syntology's graph 2026-09-24; that is when this build read the record, not when the samples ran.

2ran · our draft was wrong

Licence: 2 of the 2 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from otroshi/edgeface. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

add_padding otroshi/edgeface/face_alignment/align.py official repository ran · our draft was wrong licence not identified · pointer only · 0e38023e3f592a92 · report
replace_linear_with_lowrank_2 otroshi/edgeface/backbones/timmfr.py official repository ran · our draft was wrong licence not identified · pointer only · 94eb386a1a14f58f · report

Tasks

Face RecognitionLightweight Face Recognitionmodel

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Face Recognition CFP-FP EdgeFace - S (g=0.5) Accuracy 0.9581 #6 of 8 Archive leaderboard report
Face Recognition LFW EdgeFace - S (g=0.5) Accuracy 0.9978 #8 of 16 Archive leaderboard report
Face Recognition LFW EdgeFace - XS (g=0.6) Accuracy 0.9973 #11 of 16 Archive leaderboard report
Lightweight Face Recognition AgeDB-30 EdgeFace - S (g=0.5) Accuracy 0.9693 #1 of 5 Archive leaderboard report
Lightweight Face Recognition AgeDB-30 EdgeFace - S (g=0.5) MFLOPs 306.11 #1 of 5 Archive leaderboard report
Lightweight Face Recognition AgeDB-30 EdgeFace - S (g=0.5) MParams 3.65 #1 of 5 Archive leaderboard report
Lightweight Face Recognition AgeDB-30 EdgeFace - XS (g=0.6) Accuracy 0.96 #4 of 5 Archive leaderboard report
Lightweight Face Recognition AgeDB-30 EdgeFace - XS (g=0.6) MFLOPs 154 #4 of 5 Archive leaderboard report
Lightweight Face Recognition AgeDB-30 EdgeFace - XS (g=0.6) MParams 1.77 #4 of 5 Archive leaderboard report
Lightweight Face Recognition CALFW EdgeFace - S (g=0.5) Accuracy 0.9571 #1 of 3 Archive leaderboard report
Lightweight Face Recognition CALFW EdgeFace - S (g=0.5) MFLOPs 306.11 #1 of 3 Archive leaderboard report
Lightweight Face Recognition CALFW EdgeFace - S (g=0.5) MParams 3.65 #1 of 3 Archive leaderboard report
Lightweight Face Recognition CALFW EdgeFace - XS (g=0.6) Accuracy 0.9528 #3 of 3 Archive leaderboard report
Lightweight Face Recognition CALFW EdgeFace - XS (g=0.6) MFLOPs 154 #3 of 3 Archive leaderboard report
Lightweight Face Recognition CALFW EdgeFace - XS (g=0.6) MParams 1.77 #3 of 3 Archive leaderboard report
Lightweight Face Recognition CFP-FP EdgeFace - S (g=0.5) Accuracy 0.9581 #1 of 4 Archive leaderboard report
Lightweight Face Recognition CFP-FP EdgeFace - S (g=0.5) MFLOPs 306.11 #1 of 4 Archive leaderboard report
Lightweight Face Recognition CFP-FP EdgeFace - S (g=0.5) MParams 3.65 #1 of 4 Archive leaderboard report
Lightweight Face Recognition CFP-FP EdgeFace - XS (g=0.6) Accuracy 0.9437 #2 of 4 Archive leaderboard report
Lightweight Face Recognition CFP-FP EdgeFace - XS (g=0.6) MFLOPs 154 #2 of 4 Archive leaderboard report
Lightweight Face Recognition CFP-FP EdgeFace - XS (g=0.6) MParams 1.77 #2 of 4 Archive leaderboard report
Lightweight Face Recognition CPLFW EdgeFace - S (g=0.5) Accuracy 0.9256 #1 of 3 Archive leaderboard report
Lightweight Face Recognition CPLFW EdgeFace - S (g=0.5) MFLOPs 306.11 #1 of 3 Archive leaderboard report
Lightweight Face Recognition CPLFW EdgeFace - S (g=0.5) MParams 3.65 #1 of 3 Archive leaderboard report
Lightweight Face Recognition CPLFW EdgeFace - XS (g=0.6) Accuracy 0.9182 #2 of 3 Archive leaderboard report
Lightweight Face Recognition CPLFW EdgeFace - XS (g=0.6) MFLOPs 154 #2 of 3 Archive leaderboard report
Lightweight Face Recognition CPLFW EdgeFace - XS (g=0.6) MParams 1.77 #2 of 3 Archive leaderboard report
Lightweight Face Recognition IJB-B EdgeFace - S (g=0.5) MFLOPs 306.11 #1 of 3 Archive leaderboard report
Lightweight Face Recognition IJB-B EdgeFace - S (g=0.5) MParams 3.65 #1 of 3 Archive leaderboard report
Lightweight Face Recognition IJB-B EdgeFace - S (g=0.5) TAR @ FAR=0.01 0.9358 #1 of 3 Archive leaderboard report
Lightweight Face Recognition IJB-B EdgeFace - XS (g=0.6) MFLOPs 154 #2 of 3 Archive leaderboard report
Lightweight Face Recognition IJB-B EdgeFace - XS (g=0.6) MParams 1.77 #2 of 3 Archive leaderboard report
Lightweight Face Recognition IJB-B EdgeFace - XS (g=0.6) TAR @ FAR=0.01 0.9267 #2 of 3 Archive leaderboard report
Lightweight Face Recognition IJB-C EdgeFace - S (g=0.5) MFLOPs 306.11 #1 of 3 Archive leaderboard report
Lightweight Face Recognition IJB-C EdgeFace - S (g=0.5) MParams 3.65 #1 of 3 Archive leaderboard report
Lightweight Face Recognition IJB-C EdgeFace - S (g=0.5) TAR @ FAR=0.01 0.9563 #1 of 3 Archive leaderboard report
Lightweight Face Recognition IJB-C EdgeFace - XS (g=0.6) MFLOPs 154 #2 of 3 Archive leaderboard report
Lightweight Face Recognition IJB-C EdgeFace - XS (g=0.6) MParams 1.77 #2 of 3 Archive leaderboard report
Lightweight Face Recognition IJB-C EdgeFace - XS (g=0.6) TAR @ FAR=0.01 0.9485 #2 of 3 Archive leaderboard report
Lightweight Face Recognition LFW EdgeFace - S (g=0.5) Accuracy 0.9978 #1 of 6 Archive leaderboard report
Lightweight Face Recognition LFW EdgeFace - S (g=0.5) MFLOPs 306.11 #1 of 6 Archive leaderboard report
Lightweight Face Recognition LFW EdgeFace - S (g=0.5) MParams 3.65 #1 of 6 Archive leaderboard report
Lightweight Face Recognition LFW EdgeFace - XS (g=0.6) Accuracy 0.9973 #2 of 6 Archive leaderboard report
Lightweight Face Recognition LFW EdgeFace - XS (g=0.6) MFLOPs 154 #2 of 6 Archive leaderboard report
Lightweight Face Recognition LFW EdgeFace - XS (g=0.6) MParams 1.77 #2 of 6 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

Absolute Position EncodingsAdamAttentionBPEDense ConnectionsDropoutLabel SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSoftmaxTransformer

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