Overview
Marine resources account for 71% of the Earth's surface and have become a core strategic focus for global development. Underwater image enhancement and 3D scene reconstruction are key technologies to unlock ocean information, supporting critical applications such as marine scientific research, resource exploration, underwater archaeology, and autonomous underwater vehicle (AUV) navigation. However, extreme underwater optical environments (e.g., wavelength-dependent light attenuation, scattering, and turbidity) cause severe image degradation—including color distortion (90% red light attenuation at 10m depth), detail blurring, and low signal-to-noise ratio—while 3D reconstruction faces challenges of geometric artifacts and low fidelity in scattering-dominated scenes.
This Special Session aims to provide an interdisciplinary forum for researchers, engineers, and practitioners to share cutting-edge advances, discuss technical challenges, and explore future directions in underwater image enhancement and 3D scene reconstruction. We welcome original research papers addressing theoretical innovations, algorithm development, system implementation, and real-world applications, fostering cross-domain collaboration between multimedia processing, computer vision, marine engineering, and environmental science.
Scope & Topics
We invite submissions of high-quality papers (regular and short) on all aspects of underwater image enhancement and 3D scene reconstruction, including but not limited to:
1. Underwater Image Enhancement Technologies
- Physical model-based correction (e.g., attenuation/scattering compensation, color balance)
- Multi-modal/multi-scale fusion enhancement (optical, sonar, multi-spectral data integration)
- Deep learning-based methods (lightweight networks, generative models, transformer architectures)
- Real-time enhancement for resource-constrained devices (AUVs, underwater cameras)
- Noise suppression and artifact removal in turbid water environments
- Quantitative evaluation metrics (PSNR, SSIM, color fidelity, perceptual quality)
2. Underwater 3D Scene Reconstruction
- Physics-aware 3D reconstruction frameworks (e.g., 3D Gaussian Splatting, NeRF extensions)
- Geometric artifact reduction in scattering media (floating Gaussian pruning, uncertainty modeling)
- Multi-view stereo (MVS) for underwater environments
- Dynamic underwater scene reconstruction (4D representation of moving objects/flows)
- Cross-modal reconstruction (sonar-optical fusion, multi-sensor integration)
- High-precision mapping for marine engineering and archaeology
3. Applications & System Implementations
- Marine resource exploration and environmental monitoring
- Underwater target detection/recognition with enhanced imagery
- Autonomous navigation and obstacle avoidance for underwater vehicles
- Underwater archaeology and cultural heritage preservation
- Oceanographic survey and seabed topographic mapping
- Practical system prototypes and field test validations
4. Challenges & Emerging Directions
- Extreme environment adaptation (deep sea, high turbidity, low light)
- Cross-scene generalization and robustness optimization
- Low-data and few-shot learning for underwater vision
- Energy-efficient computing for long-duration underwater missions
- Standard datasets and benchmarking for algorithm evaluation
Important Dates
- Workshop paper submission: 25 March 2026
- Workshop paper notification: 25 April 2026
- Workshop paper camera ready: 15 May 2026
Submission Guidelines
- All submissions must be original work and not previously published or under review elsewhere
- Papers should adhere to ICME 2026's formatting requirements (double-column, IEEE Computer Society Press style).
- Submissions will be peer-reviewed by at least three experts in the field based on technical quality, originality, significance, and clarity.
- Accepted papers will be included in the ICME 2026 conference proceedings (indexed by IEEE Xplore, EI Compendex, and Scopus).
- Submit your paper via the ICME 2026 official submission system, and select the Special Session track: "Marine Underwater Information Processing and 3D Scene Reconstruction".
Workshop chairs
Liquan Shen, Shanghai University, jsslq@163.com
Guangtao Zhai, Shanghai Jiaotong University, zhaiguangtao@sjtu.edu.cn
Zhengyong Wang, Shanghai Maritime University, zy.wang@shmtu.edu.cn
Contact Information
For inquiries about the Special Session, please contact the Chairs via: jsslq@163.com
For general conference information, visit the ICME 2026 official website: https://2026.ieeeicme.org/