{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/edgeface-efficient-face-recognition-model-for","title":"EdgeFace: Efficient Face Recognition Model for Edge Devices","arxiv_id":"2307.01838","date":"2023-07-04","proceeding":null,"authors":["Anjith George","Christophe Ecabert","Hatef Otroshi Shahreza","Ketan Kotwal","Sebastien Marcel"],"abstract":"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.","url_abs":"https://arxiv.org/abs/2307.01838v2","url_pdf":"https://arxiv.org/pdf/2307.01838v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"edgeface-efficient-face-recognition-model-for","repo_url":"https://github.com/otroshi/edgeface","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"edgeface-efficient-face-recognition-model-for","repo_url":"https://github.com/anjith2006/bob.paper.tbiom2023_edgeface","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"edgeface-efficient-face-recognition-model-for","repo_url":"https://gitlab.idiap.ch/bob/bob.paper.tbiom2023_edgeface","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"face-recognition","task_name":"Face Recognition"},{"task_slug":"lightweight-face-recognition","task_name":"Lightweight Face Recognition"},{"task_slug":"model","task_name":"model"}],"methods":[{"method_slug":"absolute-position-encodings","method_name":"Absolute Position Encodings"},{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"label-smoothing","method_name":"Label Smoothing"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"position-wise-feed-forward-layer","method_name":"Position-Wise Feed-Forward Layer"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"transformer","method_name":"Transformer"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/face-recognition-on-cfp-fp","task":"Face Recognition","dataset":"CFP-FP","model":"EdgeFace - S (g=0.5)","rank_in_archive_order":6,"of":8,"metrics":{"Accuracy":"0.9581"},"uses_additional_data":false},{"leaderboard":"/sota/face-recognition-on-lfw","task":"Face Recognition","dataset":"LFW","model":"EdgeFace - S (g=0.5)","rank_in_archive_order":8,"of":16,"metrics":{"Accuracy":"0.9978"},"uses_additional_data":true},{"leaderboard":"/sota/face-recognition-on-lfw","task":"Face Recognition","dataset":"LFW","model":"EdgeFace - XS (g=0.6)","rank_in_archive_order":11,"of":16,"metrics":{"Accuracy":"0.9973"},"uses_additional_data":false},{"leaderboard":"/sota/lightweight-face-recognition-on-agedb-30","task":"Lightweight Face Recognition","dataset":"AgeDB-30","model":"EdgeFace - S (g=0.5)","rank_in_archive_order":1,"of":5,"metrics":{"Accuracy":"0.9693","MFLOPs":"306.11 ","MParams":"3.65"},"uses_additional_data":false},{"leaderboard":"/sota/lightweight-face-recognition-on-agedb-30","task":"Lightweight Face Recognition","dataset":"AgeDB-30","model":"EdgeFace - XS (g=0.6)","rank_in_archive_order":4,"of":5,"metrics":{"Accuracy":"0.96","MFLOPs":"154","MParams":"1.77"},"uses_additional_data":false},{"leaderboard":"/sota/lightweight-face-recognition-on-calfw","task":"Lightweight Face Recognition","dataset":"CALFW","model":"EdgeFace - S (g=0.5)","rank_in_archive_order":1,"of":3,"metrics":{"Accuracy":"0.9571","MFLOPs":"306.11","MParams":"3.65"},"uses_additional_data":false},{"leaderboard":"/sota/lightweight-face-recognition-on-calfw","task":"Lightweight Face Recognition","dataset":"CALFW","model":"EdgeFace - XS (g=0.6)","rank_in_archive_order":3,"of":3,"metrics":{"Accuracy":"0.9528","MFLOPs":"154","MParams":"1.77"},"uses_additional_data":false},{"leaderboard":"/sota/lightweight-face-recognition-on-cfp-fp","task":"Lightweight Face Recognition","dataset":"CFP-FP","model":"EdgeFace - S (g=0.5)","rank_in_archive_order":1,"of":4,"metrics":{"Accuracy":"0.9581","MFLOPs":"306.11","MParams":"3.65"},"uses_additional_data":false},{"leaderboard":"/sota/lightweight-face-recognition-on-cfp-fp","task":"Lightweight Face Recognition","dataset":"CFP-FP","model":"EdgeFace - XS (g=0.6)","rank_in_archive_order":2,"of":4,"metrics":{"Accuracy":"0.9437","MFLOPs":"154","MParams":"1.77"},"uses_additional_data":false},{"leaderboard":"/sota/lightweight-face-recognition-on-cplfw","task":"Lightweight Face Recognition","dataset":"CPLFW","model":"EdgeFace - S (g=0.5)","rank_in_archive_order":1,"of":3,"metrics":{"Accuracy":"0.9256","MFLOPs":"306.11","MParams":"3.65"},"uses_additional_data":false},{"leaderboard":"/sota/lightweight-face-recognition-on-cplfw","task":"Lightweight Face Recognition","dataset":"CPLFW","model":"EdgeFace - XS (g=0.6)","rank_in_archive_order":2,"of":3,"metrics":{"Accuracy":"0.9182","MFLOPs":"154","MParams":"1.77"},"uses_additional_data":false},{"leaderboard":"/sota/lightweight-face-recognition-on-ijb-b","task":"Lightweight Face Recognition","dataset":"IJB-B","model":"EdgeFace - S (g=0.5)","rank_in_archive_order":1,"of":3,"metrics":{"MFLOPs":"306.11","MParams":"3.65","TAR @ FAR=0.01":"0.9358"},"uses_additional_data":false},{"leaderboard":"/sota/lightweight-face-recognition-on-ijb-b","task":"Lightweight Face Recognition","dataset":"IJB-B","model":"EdgeFace - XS (g=0.6)","rank_in_archive_order":2,"of":3,"metrics":{"MFLOPs":"154","MParams":"1.77","TAR @ FAR=0.01":"0.9267"},"uses_additional_data":false},{"leaderboard":"/sota/lightweight-face-recognition-on-ijb-c","task":"Lightweight Face Recognition","dataset":"IJB-C","model":"EdgeFace - S (g=0.5)","rank_in_archive_order":1,"of":3,"metrics":{"MFLOPs":"306.11","MParams":"3.65","TAR @ FAR=0.01":"0.9563"},"uses_additional_data":false},{"leaderboard":"/sota/lightweight-face-recognition-on-ijb-c","task":"Lightweight Face Recognition","dataset":"IJB-C","model":"EdgeFace - XS (g=0.6)","rank_in_archive_order":2,"of":3,"metrics":{"MFLOPs":"154","MParams":"1.77","TAR @ FAR=0.01":"0.9485"},"uses_additional_data":false},{"leaderboard":"/sota/lightweight-face-recognition-on-lfw","task":"Lightweight Face Recognition","dataset":"LFW","model":"EdgeFace - S (g=0.5)","rank_in_archive_order":1,"of":6,"metrics":{"Accuracy":"0.9978","MFLOPs":"306.11","MParams":"3.65"},"uses_additional_data":false},{"leaderboard":"/sota/lightweight-face-recognition-on-lfw","task":"Lightweight Face Recognition","dataset":"LFW","model":"EdgeFace - XS (g=0.6)","rank_in_archive_order":2,"of":6,"metrics":{"Accuracy":"0.9973","MFLOPs":"154","MParams":"1.77"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2307.01838","atlas_url":"https://app.syntology.ai/?focus=2307.01838","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2307.01838"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://gitlab.idiap.ch/bob/bob.paper.tbiom2023_edgeface","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/otroshi/edgeface","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/anjith2006/bob.paper.tbiom2023_edgeface","reach":{"status":"ok"}}],"summary":{"ran_draft_wrong":2},"by_repo_kind":{"official":{"samples":2,"ran":2,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"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"},"n_pointer_only_for_licence":2,"samples":[{"code_sha256_prefix":"0e38023e3f592a92","entry":"add_padding","repo":"otroshi/edgeface","repo_kind":"official","path":"face_alignment/align.py","file_url":"https://github.com/otroshi/edgeface/blob/HEAD/face_alignment/align.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"well_formed","behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"0e38023e3f592a92"}},{"code_sha256_prefix":"94eb386a1a14f58f","entry":"replace_linear_with_lowrank_2","repo":"otroshi/edgeface","repo_kind":"official","path":"backbones/timmfr.py","file_url":"https://github.com/otroshi/edgeface/blob/HEAD/backbones/timmfr.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"94eb386a1a14f58f"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}