Research map / Detection and segmentation / CNN features
PhysFlowNet: Learning Canonical Latent Manifolds via Spatio-Spectral Physics Priors for Underwater Object Detection
XUETING Liu, Haoyu Ji, Wenze Huang, Zhihao Yang, Yu Gao, Weihong Ren, Zhiyong Wang, Honghai Liu
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 PMD (physics-guided downsampling) (ablation) | YOLO11 baseline · TrashCan | AP50: 61.9 -> 63.0 | Improved |
| add PMD (physics-guided downsampling) (ablation) | YOLO11 baseline · TrashCan | AP: 45.1 -> 46.4 | Improved |
| add HPCE physical context embedding (ablation) | YOLO11 + PMD · TrashCan | AP50: 63.0 -> 63.3 | Improved |
| add HPCE physical context embedding (ablation) | YOLO11 + PMD · TrashCan | AP: 46.4 -> 46.8 | Improved |
| add PPRB with Euclidean update (G = I) (ablation) | YOLO11 + PMD + HPCE · TrashCan | AP50: 63.3 -> 63.6 | Improved |
| use Riemannian diagonal-metric retraction instead of Euclidean update (ablation) | PPRB · TrashCan | AP50: 63.6 -> 64.9 | Improved |
| use Riemannian diagonal-metric retraction instead of Euclidean update (ablation) | PPRB · TrashCan | AP: 47.2 -> 48.4 | Improved |
| add manifold contrastive loss L_mani (ablation) | PhysFlowNet · TrashCan | AP50: 64.9 -> 65.1 | Improved |
| add evidential loss L_edl (ablation) | PhysFlowNet · TrashCan | AP50: 65.1 -> 65.4 | Improved |
| add all components to baseline (full PhysFlowNet) (ablation) | YOLO11 · TrashCan | AP: 45.1 -> 49.2 | Improved |
Closest papers by meaning
- RHCNet: Residual-Guided Hierarchical Calibration Network for Robust Underwater Object Detection (CVPR 2026 · similarity 0.87)
- WaterDiffusion: Learning a Prior-involved Unrolling Diffusion for Joint Underwater Saliency Detection and Visual Restoration (AAAI 2025 · similarity 0.86)
- WaterDescatterGS: Diverse Underwater 3D Scene Reconstruction Using Gaussian Splatting (NeurIPS 2026 · similarity 0.86)
- PhysAug: A Physical-guided and Frequency-based Data Augmentation for Single-Domain Generalized Object Detection (AAAI 2025 · similarity 0.84)
- AMSP-UOD: When Vortex Convolution and Stochastic Perturbation Meet Underwater Object Detection (AAAI 2024 · similarity 0.84)
- Task-Induced Riemannian Metrics for Vision Transformer Feature Spaces (NeurIPS 2026 · 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.
