Research map / Detection and segmentation / CNN features
Exploring the Bounds of the Utility of Context for Object Detection
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 the best object-relation context to detector scores | Faster R-CNN · COCO val, hot-dog and suitcase | AP improvement upper bound: 4.7 and 5.5 AP units | Helped |
| Add the best object-relation context to detector scores | Faster R-CNN · COCO val, zebra and cat | AP improvement upper bound: 1.5 and 1.6 AP units | Helped |
| Use co-occurrence relations without spatial-bin information | Faster R-CNN · COCO val, IoU > 0.5 | Categories whose best relation uses co-occurrence only: 70% | Helped |
| Increase the localization criterion from IoU > 0.5 to IoU > 0.75 | Object-relation context · COCO val | AP improvement upper bound | Hurt |
| Increase the localization criterion from IoU > 0.5 to IoU > 0.75 | Object-relation context · KITTI, pedestrians and cyclists | AP improvement upper bound: 1.8 -> 1; 1.6 -> 1.2 | Hurt |
| Use object relations to classify localization errors as false detections | SSD · KITTI, cars, pedestrians, and cyclists | Context classification accuracy: 57%, 55%, 54% | Hurt |
| Use object relations to classify background confusions as false detections | SSD · KITTI, cars, pedestrians, and cyclists | Context classification accuracy: 65%, 61%, 64% | Helped |
| Use object relations to classify category confusions as false detections | SSD · KITTI, cars, pedestrians, and cyclists | Context classification accuracy: 75%, 67%, 58% | Helped |
Closest papers by meaning
- ContextShift: A Controlled Benchmark for Context Dependence in Object Detection (NeurIPS 2026 · similarity 0.91)
- Size-invariance Matters: Rethinking Metrics and Losses for Imbalanced Multi-object Salient Object Detection (ICML 2024 · similarity 0.87)
- ContextHOI: Spatial Context Learning for Human-Object Interaction Detection (AAAI 2025 · similarity 0.86)
- Object Detection Benchmarks are Incomplete: The Role of Label Errors and Annotation Uncertainty (NeurIPS 2026 · similarity 0.86)
- How to Evaluate the Generalization of Detection? A Benchmark for Comprehensive Open-Vocabulary Detection (AAAI 2024 · similarity 0.85)
- DoGA: Enhancing Grounded Object Detection via Grouped Pre-Training with Attributes (AAAI 2025 · similarity 0.85)
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.