MV2Cyl: Reconstructing 3D Extrusion Cylinders from Multi-View Images

TL;DR

MV2Cyl is a new method that reconstructs 3D extrusion cylinders from multi-view images by combining surface and base curve information, leading to more accurate CAD models than previous point cloud-based approaches.

Contribution

The paper introduces MV2Cyl, a novel multi-view image-based approach for reconstructing 3D extrusion cylinders as CAD models, improving accuracy by integrating surface and base curve data.

Findings

Outperforms previous point cloud methods in accuracy

Effectively leverages multi-view images for 3D reconstruction

Achieves optimal surface and parameter estimation

Abstract

We present MV2Cyl, a novel method for reconstructing 3D from 2D multi-view images, not merely as a field or raw geometry but as a sketch-extrude CAD model. Extracting extrusion cylinders from raw 3D geometry has been extensively researched in computer vision, while the processing of 3D data through neural networks has remained a bottleneck. Since 3D scans are generally accompanied by multi-view images, leveraging 2D convolutional neural networks allows these images to be exploited as a rich source for extracting extrusion cylinder information. However, we observe that extracting only the surface information of the extrudes and utilizing it results in suboptimal outcomes due to the challenges in the occlusion and surface segmentation. By synergizing with the extracted base curve information, we achieve the optimal reconstruction result with the best accuracy in 2D sketch and extrude parameter estimation. Our experiments, comparing our method with previous work that takes a raw 3D point cloud as input, demonstrate the effectiveness of our approach by taking advantage of multi-view images. Our project page can be found at http://mv2cyl.github.io .