atlas

Research map / Detection and segmentation / CNN features

PixelLink: Detecting Scene Text via Instance Segmentation

AAAI 2018 · Detection and segmentation › CNN features · approach: Arbitrary-shape scene text detection · 652 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (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
add pixel-link predictions to group pixels into text instances (ablation)PixelLink+VGG16 4s · IC15F-score: 64.0 -> 82.3Helped
use Instance Balance in the pixel loss (ablation)PixelLink+VGG16 4s · IC15F-score: 81.2 -> 82.3Helped
use Instance Balance in the pixel loss (ablation)PixelLink+VGG16 4s · IC13F-score: 81.7 -> 83.3Helped
train using the full image size rather than 384x384 (ablation)PixelLink+VGG16 4s · IC15F-score: 80.4 -> 82.3Helped
apply post-filtering to detected boxes (ablation)PixelLink+VGG16 4s · IC15F-score: 64.3 -> 82.3Helped
predict at 2s instead of 4s resolution (ablation)PixelLink+VGG16 · IC15F-score: 82.3 -> 83.7Helped
use multi-scale testingPixelLink+VGG16 2s · IC13F-score: 84.5 -> 88.1Helped
use multi-scale testingPixelLink+VGG16 4s · IC13F-score: 83.3 -> 87.5Helped
use PixelLink instead of EASTscene text detection · COCO-Text evaluation setF-score: 39.5 -> 42.4Helped

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