AI & ML New Capability

Restores editable, semantically layered structures from flattened vector graphics (SVGs/icons) by using generative completion to recover occluded geometries.

March 26, 2026

Original Paper

SemLayer: Semantic-aware Generative Segmentation and Layer Construction for Abstract Icons

Haiyang Xu, Ronghuan Wu, Li-Yi Wei, Nanxuan Zhao, Chenxi Liu, Cuong Nguyen, Zhuowen Tu, Zhaowen Wang

arXiv · 2603.24039

The Takeaway

Practitioners frequently struggle with 'flattened' vector assets that are impossible to restyle or animate; this framework automates the decomposition of these assets back into semantic parts. It enables professional design workflows (editing, restyling, animation) on legacy or web-scraped graphics that were previously non-editable.

From the abstract

Graphic icons are a cornerstone of modern design workflows, yet they are often distributed as flattened single-path or compound-path graphics, where the original semantic layering is lost. This absence of semantic decomposition hinders downstream tasks such as editing, restyling, and animation. We formalize this problem as semantic layer construction for flattened vector art and introduce SemLayer, a visual generation empowered pipeline that restores editable layered structures. Given an abstrac