Research map / Vision-language and multimodal / Human and embodied modeling
Measuring Human Perception to Improve Open Set Recognition
Jin Huang, Derek S. Prijatelj, Justin Dulay, Walter J. Scheirer
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 exit loss to cross-entropy (ablation) | MSD-Net · ImageNet335 valid (known) | top-1 accuracy: 53.98 -> 60.00 | Improved |
| add exit loss to cross-entropy (ablation) | MSD-Net · ImageNet335 test (known) | top-1 accuracy: 26.67 -> 36.48 | Improved |
| add exit loss to cross-entropy (ablation) | MSD-Net · ImageNet335 test (unknown) | accuracy: 21.05 -> 54.23 | Improved |
| replace performance loss with exit loss (ablation) | MSD-Net CE+L_P · ImageNet335 test (known) | top-1 accuracy: 28.68 -> 36.48 | Improved |
| replace performance loss with exit loss (ablation) | MSD-Net CE+L_P · ImageNet335 test (unknown) | accuracy: 23.95 -> 54.23 | Improved |
| add performance loss to cross-entropy (ablation) | MSD-Net · ImageNet335 test (known) | top-1 accuracy: 26.67 -> 28.68 | Improved |
| add performance loss to cross-entropy (ablation) | MSD-Net · ImageNet335 test (unknown) | accuracy: 21.05 -> 23.95 | Improved |
| add performance loss to cross-entropy (ablation) | MSD-Net · ImageNet335 valid (known) | top-1 accuracy: 53.98 -> 52.90 | Hurt |
| add exit loss to cross-entropy (ablation) | MSD-Net · ImageNet335 train (known) | top-1 accuracy: 72.56 -> 72.10 | No clear effect |
| add exit loss to cross-entropy (ablation) | MSD-Net · ImageNet335 test (known) | top-5 accuracy: 41.08 -> 51.00 | Improved |
| use MSD-Net with psychophysical loss versus EVM/CROSR/CAC-OSR baselines | MSD-Net CE+exit loss · ImageNet335 test (known) | top-1 accuracy: 36.48 vs 1.92 | Improved |
Closest papers by meaning
- PsyPhy: A Psychophysics Driven Evaluation Framework for Visual Recognition (TPAMI 2018 · similarity 0.89)
- Measuring Per-Unit Interpretability at Scale Without Humans (NeurIPS 2024 · similarity 0.86)
- Evaluating Multiview Object Consistency in Humans and Image Models (NeurIPS 2024 · similarity 0.85)
- VL-SAM-V2: Open-World Object Detection with General and Specific Query Fusion (NeurIPS 2025 · similarity 0.85)
- Contour Integration Underlies Human-Like Vision (ICML 2025 · similarity 0.85)
- RTify: Aligning Deep Neural Networks with Human Behavioral Decisions (NeurIPS 2024 · similarity 0.84)
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.
