AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Learning Knowledge-Based Prompts for Robust 3D Mask Presentation Attack Detection

Fangling Jiang, Qi Li, Bing Liu, Weining Wang, Caifeng Shan, Zhenan Sun, Ming-Hsuan Yang

TPAMI 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Interleaved image-text generation with MLLMs

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
add visual-specific knowledge graph prompt (ablation)Ours · HiFiMask to 3DMAD (cross-dataset)HTER: 9.39 -> 1.04Improved
add visual-specific entity prompt (ablation)Ours · HiFiMask to 3DMAD (cross-dataset)HTER: 5.68 -> 1.04Improved
add fine-grained description prompt (ablation)Ours · HiFiMask to 3DMAD (cross-dataset)HTER: 4.30 -> 1.04Improved
add knowledge filter on entities (ablation)Ours · HiFiMask to 3DMAD (cross-dataset)HTER: 3.69 -> 1.04Improved
add knowledge filter on descriptions (ablation)Ours · HiFiMask to 3DMAD (cross-dataset)HTER: 6.37 -> 1.04Improved
add spurious correlation elimination regularization (ablation)Ours · HiFiMask to 3DMAD (cross-dataset)HTER: 2.41 -> 1.04Improved
add spurious correlation elimination regularization (ablation)Ours · HiFiMask to 3DMAD (cross-dataset)EER: 1.76 -> 1.76No clear effect
use knowledge-based prompts instead of CoOp learnable promptsOurs · HiFiMask to 3DMAD (cross-dataset)HTER: 12.39 -> 1.04Improved
use knowledge-based prompts instead of fine-tuned FLIPOurs · HiFiMask to 3DMAD (cross-dataset)HTER: 4.77 -> 1.04Improved
learn prompts instead of fixed 'A photo of' promptCoOp · HiFiMask to 3DMAD (cross-dataset)HTER: 18.10 -> 12.39Improved
scale backbone from CLIP ViT-B/16 to ViT-L/14Ours · HiFiMask to 3DMAD (cross-dataset)HTER: 1.04 -> 0.77Improved
use SigLIP ViT-B/16 backbone instead of CLIP ViT-B/16Ours · HiFiMask to 3DMAD (cross-dataset)HTER: 1.04 -> 0.86Improved

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