Research map / 3D vision / 3D generation
LSAP-PV: High-Fidelity Palm Vein Image Synthesis via Layered Spectral Absorption Projection-Guided Diffusion Model
Sheng Shang, Chenglong Zhao, Ruixin Zhang, Jianlong Jin, Jingyun Zhang, Jun Wang, Yang Zhao, Shouhong Ding, Wei Jia
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 LSAP projection (skin spectral absorption) to the generator (ablation) | LSAP-PV generator · Palm vein recognition, average of 4 datasets (4000 IDs, 7 samples/ID) | TAR@FAR=1e-4: 0.8954 -> 0.9329 | Improved |
| add vein pattern extraction module (VPEM) (ablation) | LSAP-PV generator · Palm vein recognition, average of 4 datasets (4000 IDs, 7 samples/ID) | TAR@FAR=1e-4: 0.8954 -> 0.9095 | Improved |
| add style library (ablation) | LSAP-PV generator · Palm vein recognition, average of 4 datasets (4000 IDs, 7 samples/ID) | TAR@FAR=1e-4: 0.8954 -> 0.9080 | Improved |
| combine all three components (LSAP, VPEM, style library) (ablation) | LSAP-PV generator · Palm vein recognition, average of 4 datasets (4000 IDs, 7 samples/ID) | TAR@FAR=1e-4: 0.9329 -> 0.9433 | Improved |
| increase synthetic dataset size from 1000 to 4000 IDs (ablation) | LSAP-PV synthetic data · Palm vein recognition, average of 4 datasets | TAR@FAR=1e-4: 0.9267 -> 0.9494 | Improved |
| mix real data with synthetic 4000-ID data | recognition training · Palm vein recognition, average of 4 datasets | TAR@FAR=1e-4: 0.9389 -> 0.9676 | Improved |
| train on LSAP-PV synthetic data instead of PVTree synthetic data | recognition model · Palm vein recognition, average of 4 datasets (4000 IDs, 7 samples/ID) | TAR@FAR=1e-4: 0.9202 -> 0.9433 | Improved |
| train on other synthetic palm vein or face generators instead of real data | recognition model · Palm vein recognition, average of 4 datasets | TAR@FAR=1e-4: 0.9389 (real) vs 0.5479 (NS-PVDB), 0.4107 (Syn-sPVDB) | Hurt |
Closest papers by meaning
- PVTree: Realistic and Controllable Palm Vein Generation for Recognition Tasks (AAAI 2025 · similarity 0.92)
- Diff-Palm: Realistic Palmprint Generation with Polynomial Creases and Intra-Class Variation Controllable Diffusion Models (CVPR 2025 · similarity 0.84)
- PCE-Palm: Palm Crease Energy Based Two-Stage Realistic Pseudo-Palmprint Generation (AAAI 2024 · similarity 0.83)
- VasTSD: Learning 3D Vascular Tree-state Space Diffusion Model for Angiography Synthesis (CVPR 2025 · similarity 0.79)
- vesselFM: A Foundation Model for Universal 3D Blood Vessel Segmentation (CVPR 2025 · similarity 0.79)
- RHanDS: Refining Malformed Hands for Generated Images with Decoupled Structure and Style Guidance (AAAI 2025 · similarity 0.78)
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.
