AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

NPT-Loss: Demystifying Face Recognition Losses With Nearest Proxies Triplet

Syed Safwan Khalid, Muhammad Awais, Zhen-Hua Feng, Chi-Ho Chan, Ammarah Farooq, Ali Akbari, Josef Kittler

TPAMI 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Margin-based face recognition losses · 19 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace Curricular-Face with NPT lossResNet-50 trained on CASIA · MegaFace 1M distractorsrank-1 accuracy (%): 91.52 -> 92.52Improved
replace ArcFace with NPT lossResNet-50 trained on CASIA · MegaFace 1M distractorsrank-1 accuracy (%): 89.01 -> 92.52Improved
replace naive proxy-triplet with NPT loss (implicit hard-negative mining)ResNet-50 trained on CASIA · MegaFace 1M distractorsrank-1 accuracy (%): 35.85 -> 92.52Improved
replace ArcFace with NPT lossResNet-50 trained on CASIA · CFP-FPaccuracy (%): 95.3 -> 96.44Improved
replace ArcFace with NPT lossResNet-50 trained on CASIA · LFWaccuracy (%): 99.3 -> 99.49Improved
replace Curricular-Face with NPT lossResNet-50 trained on MS1Mv2 · MegaFacerank-1 accuracy (%): 96.98 -> 97.67Improved
replace ArcFace with NPT lossResNet-50 trained on MS1Mv2 · MegaFacerank-1 accuracy (%): 97.58 -> 97.67Improved
replace ArcFace with NPT lossResNet-50 trained on MS1Mv2 · IJB-CTAR @ FAR=1e-4 (%): 94.79 -> 94.97Improved
replace ArcFace with NPT lossResNet-50 trained on MS1Mv2 · IJB-BTAR @ FAR=1e-4 (%): 93.15 -> 93.39Improved
replace Huang et al. with NPT loss on low-resolution trainingResNet-50 · SCFaceaverage accuracy (%): 94.46 -> 96.61Improved
vary margin-like hyper-parameter Delta above r^2/2 (ablation)NPT loss, CASIA · MegaFace (10K distractors)rank-1 accuracyNo clear effect
replace Curricular-Face with NPT loss at very low FARResNet-50 trained on CASIA · MegaFace ROCTARHurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy