Learning Volumetric Shape Super-Resolution for Cranial Implant Design

Matthias Eder, Jianning Li, Jan Egger*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference paperpeer-review


Cranioplasty is the process of repairing cranial defects or deformations, which may be the result of injuries or necessary medical treatments such as brain tumor surgery. For this procedure, it is necessary to generate a high-quality cranial implant, which needs to be shaped individually for each skull and each defect. This tends to be a very time consuming task and requires also in-depth knowledge of various CAM/CAD programs. In this work, we present a novel automatic three-stage implant generation pipeline. First, skull completion is conducted in low resolution using a trained artificial neural network (ANN). Second, the completed low-resolution skull is sent to a super-resolution network, which up-samples the low-resolution skull to higher resolution while, at the same time, filling the skull surface with geometric details. Finally, by simple subtraction and blob filtering, the desired high-resolution implant is generated.
Original languageEnglish
Title of host publicationTowards the Automatization of Cranial Implant Design in Cranioplasty - First Challenge, AutoImplant 2020, Held in Conjunction with MICCAI 2020, Proceedings
EditorsJianning Li, Jan Egger
PublisherSpringer, Cham
Pages104 -113
Number of pages10
ISBN (Electronic)978-3-030-64327-0
ISBN (Print)978-3-030-64326-3
Publication statusE-pub ahead of print - 4 Dec 2020
Event1st Automatization of Cranial Implant Design in Cranioplasty Challenge: AutoImplant 2020 - Virtual, Lima, Peru
Duration: 8 Oct 2020 → …

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12439 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference1st Automatization of Cranial Implant Design in Cranioplasty Challenge
CityVirtual, Lima
Period8/10/20 → …
Internet address


  • Cranioplasty
  • Deep learning
  • Shape completion
  • Super-resolution

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)


Dive into the research topics of 'Learning Volumetric Shape Super-Resolution for Cranial Implant Design'. Together they form a unique fingerprint.

Cite this