Research map / Detection and segmentation / Open-set and few-shot detection
Attention Guided Anomaly Localization in Images
Shashanka Venkataramanan, Kuan-Chuan Peng, Rajat Vikram Singh, Abhijit Mahalanobis
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 attention expansion loss L_ae (ablation) | CAVGA-R_u* (flattened z) · MVTAD (ablation, 15 categories) | mean IoU: 0.24 -> 0.34 | Improved |
| add attention expansion loss L_ae (ablation) | CAVGA-R_u (conv z) · MVTAD (ablation, 15 categories) | mean IoU: 0.33 -> 0.47 | Improved |
| add complementary guided attention loss L_cga (ablation) | CAVGA-R_w* (flattened z) · MVTAD (ablation, 15 categories) | mean IoU: 0.39 -> 0.52 | Improved |
| add complementary guided attention loss L_cga (ablation) | CAVGA-R_w (conv z) · MVTAD (ablation, 15 categories) | mean IoU: 0.48 -> 0.60 | Improved |
| use convolutional latent variable instead of flattened, without L_ae (ablation) | CAVGA-R_u · MVTAD (ablation, 15 categories) | mean IoU: 0.24 -> 0.33 | Improved |
| use convolutional latent variable instead of flattened, with L_ae (ablation) | CAVGA-R_u · MVTAD (ablation, 15 categories) | mean IoU: 0.34 -> 0.47 | Improved |
| use convolutional latent variable instead of flattened, without L_cga (ablation) | CAVGA-R_w · MVTAD (ablation, 15 categories) | mean IoU: 0.39 -> 0.48 | Improved |
| use convolutional latent variable instead of flattened, with L_cga (ablation) | CAVGA-R_w · MVTAD (ablation, 15 categories) | mean IoU: 0.52 -> 0.60 | Improved |
| add 2% anomalous images with classifier (weak supervision) | CAVGA-D · MVTAD localization | mean IoU: 0.41 -> 0.50 | Improved |
| add 2% anomalous images with classifier (weak supervision) | CAVGA-D · mSTC localization | mean IoU: 0.239 -> 0.330 | Improved |
| use ResNet-18 encoder with residual decoder instead of DC-GAN encoder/decoder | CAVGA_u · MVTAD localization | mean IoU: 0.41 -> 0.47 | Improved |
Closest papers by meaning
- Attention-Conditioned Augmentations for Self-Supervised Anomaly Detection and Localization (AAAI 2023 · similarity 0.88)
- Glancing at the Patch: Anomaly Localization With Global and Local Feature Comparison (CVPR 2021 · similarity 0.87)
- Boosting Fine-Grained Visual Anomaly Detection with Coarse-Knowledge-Aware Adversarial Learning (AAAI 2025 · similarity 0.87)
- PatchGuard: Adversarially Robust Anomaly Detection and Localization through Vision Transformers and Pseudo Anomalies (CVPR 2025 · similarity 0.86)
- ADPretrain: Advancing Industrial Anomaly Detection via Anomaly Representation Pretraining (NeurIPS 2025 · similarity 0.84)
- Local-Prompt: Extensible Local Prompts for Few-Shot Out-of-Distribution Detection (ICLR 2025 · 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.
