Learning-based reconstruction of moving transparent objects

This work focuses on learning-based reconstruction of moving transparent objects, contributing to Wei Pan’s research thread in 3D vision, optical metrology, geometry processing, and industrial inspection.

In the broader publication record, this work sits in International Conference on Optoelectronic Science and Intelligent Sensing (SPIE), 1417511 and connects to practical problems in 3D sensing, computational geometry, and industrial machine vision.

Algorithm principle

The method combines a domain-specific representation with an optimization or learning pipeline. It first converts raw visual, geometric, or optical measurements into features that expose the underlying 3D structure, then refines those features into a reconstruction, correspondence field, segmentation, measurement, or control decision.

Visual material

The local archive does not currently include a matched PDF for this entry, so additional method and result figures are pending. Once the PDF is added, the page can be regenerated with representative visuals.

Results and impact

The reported evaluation in International Conference on Optoelectronic Science and Intelligent Sensing (SPIE), 1417511 positions the method as a practical contribution to the surrounding 3D vision and industrial inspection workflow. The experiments are used to show whether the proposed representation or pipeline improves robustness, accuracy, speed, or deployability compared with the relevant baseline methods.