Research map / Detection and segmentation / CNN features
Deployment Prior Injection for Run-Time Re-Biasable Object Detection
Mo Zhou, Yiding Yang, Haoxiang Li, Vishal M. Patel, Gang Hua
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 |
|---|---|---|---|
| inject accurate per-sample prior E_x instead of default E_t | CaliDet (DINO) · COCO val | AP: 49.27 -> 51.94 | Improved |
| inject mini-batch prior E_b instead of default E_t | CaliDet (DINO) · COCO val | AP: 49.27 -> 51.69 | Improved |
| inject flat prior E_0 instead of E_t | CaliDet (DINO) · COCO val | AP: 49.27 -> 47.97 | Hurt |
| inject misleading prior instead of E_t | CaliDet (DINO) · COCO val | AP: 49.27 -> 35.83 | Hurt |
| add run-time calibration (default prior) to DINO | DINO · COCO val | AP: 49.07 -> 49.27 | No clear effect |
| inject subset statistics E_b(8) instead of E_t | CaliDet (DINO) · COCO val subsets (size 8) | AP: 62.93 -> 63.15 | Improved |
| inject subset statistics E_b(256) instead of E_t | CaliDet (DINO) · COCO val subsets (size 256) | AP: 53.67 -> 53.82 | Improved |
| inject per-sample prior E_x instead of E_t | CaliDet (D-DETR) · COCO val | AP: 41.69 -> 43.92 | Improved |
| inject subset statistics E_b(8) instead of E_t | CaliDet (D-DETR) · COCO val subsets (size 8) | AP: 55.35 -> 55.64 | Improved |
| add Logit Manipulation loss (ablation) | CaliDet · COCO val | AP | Improved |
| add Gaussian noise to sampled edges during training (ablation) | CaliDet · COCO val | AP | Improved |
| sample priors from E_x, E_b, E_t instead of only E_t or E_x (ablation) | CaliDet · COCO val | AP | Improved |
Closest papers by meaning
- ContextShift: A Controlled Benchmark for Context Dependence in Object Detection (NeurIPS 2026 · similarity 0.85)
- Detecting Objects with Context-Likelihood Graphs and Graph Refinement (ICCV 2023 · similarity 0.85)
- Long-tailed Object Detection Pretraining: Dynamic Rebalancing Contrastive Learning with Dual Reconstruction (NeurIPS 2024 · similarity 0.83)
- What How and When Should Object Detectors Update in Continually Changing Test Domains? (CVPR 2024 · similarity 0.83)
- Interpreting Object-level Foundation Models via Visual Precision Search (CVPR 2025 · similarity 0.83)
- Test-Time Backdoor Detection for Object Detection Models (CVPR 2025 · similarity 0.83)
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.
