Gaussian Splatting with Reflectance Regularization for Endoscopic Scene Reconstruction

Endoscopic reconstruction plays a crucial role in surgical robotics. The dynamic lighting conditions and integrated camera-light source in endoscopic scenes create a distinct reconstruction challenge: shape ambiguity. To mitigate this, we propose a Gaussian Splatting (GS) based framework for endoscopic scene reconstruction, enhanced with reflectance regularization. We embed every 3D Gaussian point with physical reflective attributes and combine this representation with a physically based inverse rendering framework. By jointly training 3DGS for view synthesis with this reflectance regularization, we are able to attain high-quality geometry without changing the volume rendering pipeline. Our experiments demonstrate the superiority in both geometry representation and rendering performance compared to existing GS approaches, making it a practical solution for endoscopic applications. Project is available at: https://med-air.github.io/GSR2.

Authors

• Chengkun Li
• Kai Chen 0028
• Shi Qiu
• Jason Ying-Kuen Chan
• Qi Dou 0001

Keywords

• Reflectivity
• Geometry
• Training
• Three-dimensional displays
• Medical robotics
• Shape
• Pipelines
• Lighting
• Rendering (computer graphics)
• Intelligent robots
• Endoscopic
• Physical Properties
• 3D Point
• Representation Of Geometry
• 3D Gaussian
• Gauss Points
• View Synthesis
• Root Mean Square Error
• Light Source
• Optimal Parameters
• Point Cloud
• Colonoscopy
• Diffuse Reflectance
• Depth Map
• Peak Signal-to-noise Ratio
• Illumination Conditions
• Normal Approximation
• Depth Estimation
• 3D Scene
• Postoperative Assessment
• Structural Similarity Index Measure
• Surgical Tasks
• Geometric Representation
• Geometric Consistency
• Normal Map
• View Direction
• Depth Error
• Geometric Information
• Direction Of The Incident Light
• Normal Error