Research map / Detection and segmentation / Open-set and few-shot detection
Semantic Novelty Detection via Relational Reasoning
Francesco Cappio Borlino, Silvia Bucci, Tatiana Tommasi
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 |
|---|---|---|---|
| use learned transformer relational module instead of Max aggregation (ablation) | ReSeND · PACS multi-source | avg AUROC: 0.726 -> 0.765 | Improved |
| use learned transformer relational module instead of Sum aggregation (ablation) | ReSeND · PACS multi-source | avg AUROC: 0.530 -> 0.765 | Improved |
| use learned transformer relational module instead of Concat aggregation (ablation) | ReSeND · PACS multi-source | avg AUROC: 0.736 -> 0.765 | Improved |
| train with MSE regression instead of cross-entropy classification (ablation) | ReSeND · Open-set DG benchmarks | AUROC | Improved |
| add ReSeND score to SagNet MSP | SagNet · Open-set DG single-source | AUROC: 0.643 -> 0.700 | Improved |
| add ReSeND score to Diversify MSP | Diversify · Open-set DG single-source | AUROC: 0.643 -> 0.691 | Improved |
| add ReSeND score to DAML MSP | DAML · Open-set DG multi-source | AUROC: 0.657 -> 0.722 | Improved |
| add ReSeND score to SWAD MSP | SWAD · Open-set DG multi-source | AUROC: 0.570 -> 0.700 | Improved |
| add ReSeND score to SWAD MSP | SWAD · Open-set DG multi-source | H-score: 42.85 -> 57.05 | Improved |
| add ReSeND score to DAML MSP | DAML · Open-set DG multi-source | H-score: 52.99 -> 57.93 | Improved |
| finetune Mahalanobis on the support set | Mahalanobis · PACS multi-source | avg AUROC: 0.616 -> 0.665 | Improved |
| finetune Gram on the support set | Gram · PACS multi-source | avg AUROC: 0.656 -> 0.677 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CutMix — CutMix: Regularization Strategy to Train Strong Classifiers With Localizable Features (ICCV 2019) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Transfer-Based Semantic Anomaly Detection (ICML 2021 · similarity 0.86)
- Detecting Semantic Anomalies (AAAI 2020 · similarity 0.85)
- Semantic-Guided Novel Category Discovery (AAAI 2024 · similarity 0.84)
- SetAD: Semi-Supervised Anomaly Learning in Contextual Sets (NeurIPS 2026 · similarity 0.82)
- SEAL: Semantic-Aware Hierarchical Learning for Generalized Category Discovery (NeurIPS 2025 · similarity 0.82)
- Open-Vocabulary Video Anomaly Detection (CVPR 2024 · similarity 0.82)
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.
