AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

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

CVPR 2026 · Detection and segmentation › Open-set and few-shot detection · approach: Object-level representation learning · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use colorization-based generative segmentation instead of finetuned SAM2*SemLayer segmentation · Icon test set (48 icons)PQ: 59.4 -> 76.1Improved
use generative segmentation instead of finetuned SAM2*SemLayer segmentation · Icon test set (48 icons)mIoU: 79.3 -> 84.3Improved
use generative segmentation instead of off-the-shelf SAM2SemLayer segmentation · Icon test set (48 icons)mIoU: 51.1 -> 84.3Improved
apply nearest-neighbor refinement of unlabeled pixelsSemLayer segmentation · Icon test set (48 icons)PQ: 76.1 -> 78.3Improved
apply nearest-neighbor refinementSAM2* · Icon test set (48 icons)PQ: 59.4 -> 78.0Improved
apply nearest-neighbor refinementgpt-image-1 · Icon test set (48 icons)mIoU: 25.4 -> 57.2Improved
use better segmentation inputs to fixed completion moduleSemLayer · Icon test set (48 icons)Chamfer Distance: 49.1 -> 46.6Improved
finetune pix2gestalt completion on icon data instead of MP3D finetunedSemLayer completion · Icon test set (48 icons)mIoU: 75.3 -> 85.2Improved
use generative diffusion priors instead of classification/regression baselineSemLayer · Icon test set (48 icons)segmentation mIoU: 46.6 -> 84.3Improved
use generative diffusion priors instead of classification/regression baselineSemLayer · Icon test set (48 icons)completion mIoU: 63.9 -> 85.2Improved
use pixel-based pipeline instead of LLMs on raw SVG codeGemini-3 and GPT-5 · Icon test set (48 icons)validityHurt
use completion model instead of MP3D for speedSemLayer completion · Icon test set (48 icons)seconds per icon: 49.27 -> 16.83Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy