Research map / Label-efficient and robust learning / Recognition and long-tail
Unconstrained Still/Video-Based Face Verification with Deep Convolutional Neural Networks
Jun-Cheng Chen, Rajeev Ranjan, Swami Sankaranarayanan, Amit Kumar, Ching‐Hui Chen, Vishal M. Patel, Carlos D. Castillo, Rama Chellappa
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add media averaging of features (ablation) | DCNN setup 3 · IJB-A verification FAR=1e-2 | TAR: 0.846 -> 0.889 | Improved |
| add media averaging of features (ablation) | DCNN setup 3 · IJB-A identification Rank-1 | accuracy: 0.922 -> 0.942 | Improved |
| add media averaging of features (ablation) | DCNN setup 3 · JANUS CS2 verification FAR=1e-2 | TAR: 0.922 -> 0.935 | Improved |
| add media averaging of features (ablation) | DCNN setup 3 · JANUS CS2 identification Rank-1 | accuracy: 0.909 -> 0.931 | Improved |
| use semi-automatic setup 3 instead of manual metadata setup 1 | DCNN_m · IJB-A verification FAR=1e-2 | TAR: 0.863 -> 0.889 | Improved |
| use automatic setup 2 instead of manual setup 1 | DCNN_m · IJB-A verification FAR=1e-2 | TAR: 0.863 -> 0.885 | Improved |
| use automatic setup 2 instead of manual setup 1 | DCNN · IJB-A identification Rank-1 | accuracy: 0.915 -> 0.907 | No clear effect |
| fuse DCNN_S and DCNN_L similarity scores | DCNN_S · LFW | accuracy: 97.7 -> 98 | Improved |
| add triplet similarity embedding to fused features (ablation) | DCNN_S + DCNN_L · LFW | accuracy: 98 -> 98.33 | Improved |
| use tight-box DCNN_S instead of large-box DCNN_L | DCNN · LFW | accuracy: 96.8 -> 97.7 | Improved |
| add fine-tuning on JANUS CS2 training data (ablation) | DCNN_S and DCNN_L · JANUS CS2 verification | TAR | Improved |
| add triplet similarity embedding (ablation) | DCNN_S and DCNN_L · JANUS CS2 verification | TAR | Improved |
Closest papers by meaning
- Joint Registration and Representation Learning for Unconstrained Face Identification (CVPR 2017 · similarity 0.88)
- Hybrid Deep Learning for Face Verification (TPAMI 2016 · similarity 0.87)
- Pre-training Strategies and Datasets for Facial Representation Learning (ECCV 2022 · similarity 0.87)
- Detecting Faces Using Inside Cascaded Contextual CNN (ICCV 2017 · similarity 0.86)
- Neural Aggregation Network for Video Face Recognition (CVPR 2017 · similarity 0.86)
- BioMetricNet: deep unconstrained face verification through learning of metrics regularized onto Gaussian distributions (ECCV 2020 · similarity 0.86)
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.
