Sketch-based modeling with perception-aware extraction and intention-aware snapping of contours
Year: 2025
Type: article
Abstract: Using 2D sketches as input is an intuitive way to create 3D models. However, accurately sketching freeform strokes by hand from a reference drawing can be time-consuming and repetitive. This paper addresses this issue by proposing an efficient sketch modeling method based on structural part extracting and intention-aware snapping of an input drawing. Based on drawing contours, perceptually completed parts are decomposed after completing occlusions, then the boundaries of parts are used as construction curves to inflate 3D shapes. Moreover, user-drawn strokes can be seamlessly snapped onto drawing contours while preserving the user’s intention, allowing users to create shapes with different semantics. This approach enables users to easily obtain desired and accurate construction curves without the need for careful contour alignment. To achieve irregular shape design, 3D shapes can be further deformed by a contour deformation method from different viewpoints. We evaluate the method on several drawings and demonstrate our method effectively facilitates the creation of desired 3D models from a guide drawing.
DOI: 10.1093/jcde/qwaf018
Authors: Yinghan Jin, Hao Huang, Jituo Li, Liuqiao Guo, Yifan Zheng, Dongliang Zhang
Institutions: Zhejiang Normal University, Zhejiang University
Language: English
Subfield: Computational Mechanics
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,650
Awards: G2284401721, 面向玩具的单视图草图建模与二维纸样生成方法研究
OpenAlex ID: W4407243892
