Research map / 3D vision / 3D generation
Safe Multi-View Deep Classification
Wei Liu, Yufei Chen, Xiaodong Yue, Changqing Zhang, Shaorong Xie
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 |
|---|---|---|---|
| Use evidence-level fusion with uncertainty and KL-divergence losses instead of the strongest compared baseline | Multi-view classification · HMDB | accuracy (%): 74.98±1.02 -> 90.84±0.11 | Improved |
| Add uncertainty loss to evidence fusion without KL-divergence loss (ablation) | SMDC · Caltech101 | accuracy (%): 96.00±0.12 -> 96.89±0.20 | Improved |
| Add KL-divergence loss to evidence fusion without uncertainty loss (ablation) | SMDC · Caltech101 | accuracy (%): 96.00±0.12 -> 97.33±0.12 | Improved |
| Add uncertainty and KL-divergence losses to evidence fusion (ablation) | SMDC · Caltech101 | accuracy (%): 96.00±0.12 -> 97.78±0.01 | Improved |
| Add view V3 to V1+V2 | SMDC · Handwritten | accuracy (%): 85.25±0.04 -> 96.75±0.01 | Improved |
| Add view V4 to V1+V2+V3 | SMDC · Handwritten | accuracy (%): 96.75±0.01 -> 97.50±0.02 | Improved |
| Add view V5 to V1+V2+V3+V4 | SMDC · Handwritten | accuracy (%): 97.50±0.02 -> 98.75±0.02 | Improved |
| Add view V6 to V1+V2+V3+V4+V5 | SMDC · Handwritten | accuracy (%): 98.75±0.02 -> 99.00±0.01 | Improved |
| Add a noisy view at noise ratio λ = 0.5 rather than using the pre-addition baseline | SMDC · CUB | accuracy (%): 74.17±0.01 -> 78.31±0.01 | Improved |
| Add a noisy view at noise ratio λ = 0.5 rather than using the pre-addition baseline | SMDC · Scene15 | accuracy (%): 56.96±1.23 -> 65.33±1.12 | Improved |
Closest papers by meaning
- Enhancing Testing-Time Robustness for Trusted Multi-View Classification in the Wild (CVPR 2025 · similarity 0.85)
- Trusted Unified Feature-Neighborhood Dynamics for Multi-View Classification (AAAI 2025 · similarity 0.84)
- Enhancing Multi-View Classification Reliability with Adaptive Rejection (AAAI 2025 · similarity 0.84)
- Deep Fuzzy Multi-view Learning for Reliable Classification (ICML 2025 · similarity 0.84)
- Trusted Multi-View Classification with Expert Knowledge Constraints (ICML 2025 · similarity 0.84)
- Investigating and Mitigating the Side Effects of Noisy Views for Self-Supervised Clustering Algorithms in Practical Multi-View Scenarios (CVPR 2024 · similarity 0.83)
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.
