Research map / Detection and segmentation / CNN features
Structure Inference Net: Object Detection Using Scene-Level Context and Instance-Level Relationships
Read the paperCheck 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 scene-context message passing (ablation) | Faster R-CNN · PASCAL VOC 2007 test | mAP: 68.79 -> 70.23 | Helped |
| add object-object relationship message passing (ablation) | Faster R-CNN · PASCAL VOC 2007 test | mAP: 68.79 -> 70.31 | Helped |
| add scene-context message passing (ablation) | Faster R-CNN · COCO 2015 test-dev | AP: 21.1 -> 22.5 | Helped |
| add object-object relationship message passing (ablation) | Faster R-CNN · COCO 2015 test-dev | AP: 21.1 -> 22.7 | Helped |
| combine scene context and object relationships using mean-pooling (ablation) | SIN · PASCAL VOC 2007 test | mAP: 70.2 -> 70.5 | Helped |
| replace mean-pooling with max-pooling to combine Scene and Edge (ablation) | SIN · PASCAL VOC 2007 test | mAP: 70.5 -> 70.4 | Hurt |
| replace mean-pooling with concatenation to combine Scene and Edge (ablation) | SIN · PASCAL VOC 2007 test | mAP: 70.5 -> 70.2 | Hurt |
| reduce message-passing updates from 2 steps to 1 (ablation) | SIN · PASCAL VOC 2007 test | mAP: 70.5 -> 69.8 | Hurt |
| increase message-passing updates from 2 steps to 3 (ablation) | SIN · PASCAL VOC 2007 test | mAP: 70.5 -> 69.6 | Hurt |
| use scene context and object relationships jointly (ablation) | Faster R-CNN · PASCAL VOC 2007 test | mAP: 68.79 -> 70.5 | Helped |
| use object-object relationship reasoning (ablation) | Faster R-CNN · COCO 2015 test-dev | AP^70: 19.9 -> 21.6 | Helped |
Closest papers by meaning
- Attentive Relational Networks for Mapping Images to Scene Graphs (CVPR 2019 · similarity 0.88)
- Spatial-Aware Graph Relation Network for Large-Scale Object Detection (CVPR 2019 · similarity 0.88)
- Spatially Invariant Unsupervised Object Detection with Convolutional Neural Networks (AAAI 2019 · similarity 0.88)
- Scene Graph Generation by Iterative Message Passing (CVPR 2017 · similarity 0.87)
- Object DGCNN: 3D Object Detection using Dynamic Graphs (NeurIPS 2021 · similarity 0.87)
- ContextShift: A Controlled Benchmark for Context Dependence in Object Detection (NeurIPS 2026 · similarity 0.86)
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.