Research map / Detection and segmentation / CNN features
Dual Refinement Underwater Object Detection Network
Baojie Fan, Wei Chen, Yang Cong, Jiandong Tian
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 composite connection block (ablation) | RFBNet baseline · UWD | mAP: 60.0 -> 63.8 | Improved |
| add focal loss (ablation) | RFBNet baseline · UWD | mAP: 60.0 -> 63.2 | Improved |
| add focal loss to composite connection (ablation) | BL+C · UWD | mAP: 63.8 -> 66.6 | Improved |
| add random erasing (ablation) | BL+C+F · UWD | mAP: 66.6 -> 66.7 | No clear effect |
| add prediction refinement scheme (PRS) (ablation) | BL+C+R+F · UWD | mAP: 66.7 -> 74.2 | Improved |
| add focal loss on top of PRS (ablation) | FERNet · UWD | mAP: 74.2 -> 73.0 | Hurt |
| add BatchNorm to VGG16 backbone (ablation) | FERNet · UWD | mAP: 74.2 -> 73.0 | Hurt |
| use six prediction layers instead of four (ablation) | FERNet · UWD | mAP: 73.3 -> 74.2 | Improved |
| use six prediction layers instead of four (ablation) | FERNet · PASCAL VOC 2007 | mAP: 79.8 -> 80.2 | Improved |
| FERNet versus RFBNet baseline | FERNet · UWD | mAP: 60.0 -> 74.2 | Improved |
| FERNet versus RFBNet on general benchmark | FERNet · PASCAL VOC 2007 (300x300) | mAP: 80.0 -> 80.2 | No clear effect |
| FERNet at 512 input versus RFBNet512 | FERNet · PASCAL VOC 2007 (512x512) | mAP: 82.1 -> 81.0 | Hurt |
Closest papers by meaning
- RHCNet: Residual-Guided Hierarchical Calibration Network for Robust Underwater Object Detection (CVPR 2026 · similarity 0.87)
- PhysFlowNet: Learning Canonical Latent Manifolds via Spatio-Spectral Physics Priors for Underwater Object Detection (ECCV 2026 · similarity 0.85)
- Synergistic Multiscale Detail Refinement via Intrinsic Supervision for Underwater Image Enhancement (AAAI 2024 · similarity 0.85)
- AMSP-UOD: When Vortex Convolution and Stochastic Perturbation Meet Underwater Object Detection (AAAI 2024 · similarity 0.84)
- WaterDiffusion: Learning a Prior-involved Unrolling Diffusion for Joint Underwater Saliency Detection and Visual Restoration (AAAI 2025 · similarity 0.84)
- Adaptive Dual-domain Learning for Underwater Image Enhancement (AAAI 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.
