AI 中文总结
介绍IJCB 2026上的AFMFR竞赛，含全数据和有限数据两个赛道，用IDPERTURB生成训练数据，基于多种基准评估。结果显示合成训练数据适配CLIP模型有改进，全数据赛道全量微调领先，有限数据条件下秩稳定的LoRA适配最有效。
AI 中文摘要
本文概述了在2026年国际生物识别联合会议（IJCB 2026）上举办的使用合成训练数据进行人脸识别基础模型适配竞赛（AFMFR）。竞赛收到来自四个不同团队的八份有效提交，分两个互补赛道：全数据赛道，参与者使用大规模合成身份数据适配CLIP ViT-L/14基础模型；有限数据赛道，反映资源受限的适配情况。所有训练数据仅使用IDPERTURB生成。提交的解决方案根据在包括LFW、CFP-FP等多种基准上的验证和识别性能，用博尔达计数法排名。还对RFW数据集的四个人口统计群体进行公平性评估。结果表明，用合成训练数据适配CLIP基础模型比现成模型有显著改进，在某些情况下超过基线。特别是，使用子中心弧脸（DMSTI-神经技术）的全量微调在全数据赛道领先，而秩稳定的LoRA适配（Idiap-BSP）在有限数据条件下最有效。
英文摘要
This paper presents a summary of the Competition on Adapting Foundation Models for Face Recognition Using Synthetic Training Data (AFMFR), held at the 2026 International Joint Conference on Biometrics (IJCB 2026). The competition received a total of eight valid submissions from four distinct teams across two complementary tracks: a Full Data Track, in which participants adapt the CLIP ViT-L/14 foundation model using large-scale synthetic identity data, and a Limited Data Track, designed to reflect more resource-constrained adaptation regimes. All training data was generated exclusively using IDPERTURB. Submitted solutions are ranked based on verification and identification performance across a diverse suite of benchmarks, including LFW, CFP-FP, AgeDB-30, CALFW, CPLFW, IJB-B, IJB-C, and TinyFace, using the Borda count method. Fairness evaluation is additionally conducted on the RFW dataset across four demographic groups. The results demonstrate that adaptation of the CLIP foundation model with synthetic training data substantially improves over the off-the-shelf model and, in several cases, surpasses the baseline. Notably, full fine-tuning with Sub-Center ArcFace (DMSTI-Neurotechnology) leads the Full Data Track, while rank-stabilized LoRA adaptation (Idiap-BSP) proves most effective under limited-data conditions.
CommentsAccepted at the IEEE International Joint Conference on Biometrics 2026 (IJCB 2026)
