AI Paper Atlas

Research map / Detection and segmentation / CNN features

Segmentation of Tubular Structures Using Iterative Training with Tailored Samples

Wei Liao

ICCV 2023 · Detection and segmentation › CNN features · approach: Context modeling for semantic segmentation · 6 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
iterative training with tailored samples instead of samples drawn directly along annotationsminimal path method with CNN · EPFL roadscenterline mean error (pixels): 1.71 -> 1.41Improved
iterative training with tailored samples vs Path-CNNminimal path method with CNN · Toronto roadscenterline mean error (pixels): 3.65 -> 3.54Improved
iterative tailored-sample training instead of Path-CNN trainingminimal path method · EPFL roadsmean Dice: 0.6322 -> 0.9125Improved
iterative tailored-sample training instead of Path-CNN trainingminimal path method · DRIVE (3 train images)mean Dice: 0.5470 -> 0.7578Improved
compare against FR-UNet on full DRIVE training setproposed method · DRIVE (20 train images)mean Dice: 0.8065 vs 0.8208Hurt
run more training iterations of the iterative scheme (ablation)proposed method · all three datasets (validation Dice)validation DiceImproved
use CNN patch features in rectified patches vs per-pixel CNN weights (Dijkstra CNN)minimal path method · EPFL roadscenterline mean error (pixels): 2.91 -> 1.41Improved

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