---

layout: pagewide

title: "Whole Body Organ Segmentation"

subheadline: ""

show_meta: false

articlename: "Whole Body Organ Segmentation"

teaser: ""

noheading: "true"

permalink: "/projects/organs/"

header:

image_fullwidth: "header_unsplash_railroad.jpg"

---

<div class="container-fluid t30">

<div class="row t30">

<h5 class="navigatorButtons"><a href="/projects/">Projects</a> / <a href="/projects/#imga">Image Analysis</a> / {{ page.articlename }} </h5>

<h1>Automatic Multi-Organ Segmentation Using Learning-based Segmentation and Level Set Optimization</h1><br>

<p class=text><b><A NAME="overview">Abstract</A></b></p>

<p class=text>We present a novel generic segmentation system for the fully automatic multi-organ segmentation from CT medical images. Thereby we combine the advantages of learning-based approaches on point cloud-based shape representation, such a speed, robustness, point correspondences, with those of PDE-optimization-based level set approaches, such as high accuracy and the straightforward prevention of segment overlaps. In a benchmark on 10-100 annotated datasets for the liver, the lungs, and the kidneys we show that the proposed system yields segmentation accuracies of 1.17-2.89mm average surface errors. Thereby the level set segmentation (which is initialized by the learning-based segmentations) contributes with an 20%-40% increase in accuracy. </p>

<p class=text><b><A NAME="overview">Publications and Further Reading</A></b></p>

{% bibliography --query @*[key ^= kohlberger:miccai11] %}

</div>