Conference paper Open Access

Probabilistic modelling combined with a CNN for boundary detection of carbon fiber fabrics

Sebastian Zambal; Christoph Heindl; Christian Eitzinger


JSON-LD (schema.org) Export

{
  "description": "<p>Abstract:</p>\n\n<p>For many industrial machine vision applications it is difficult to acquire good training data to deploy deep learning techniques. In this paper we propose a method based on probabilistic modelling and rendering to generate artificial images of carbon fiber fabrics. We deploy a convolutional neural network (CNN) to learn detection of fabric contours from artificially generated images. Our network largely follows the recently proposed U-Net architecture. We provide results for a set of real images taken under controlled lighting conditions. The method can easily be adapted to similar problems in quality control for composite parts.</p>", 
  "license": "", 
  "creator": [
    {
      "affiliation": "PROFACTOR GmbH", 
      "@type": "Person", 
      "name": "Sebastian Zambal"
    }, 
    {
      "affiliation": "PROFACTOR GmbH", 
      "@type": "Person", 
      "name": "Christoph Heindl"
    }, 
    {
      "affiliation": "PROFACTOR GmbH", 
      "@type": "Person", 
      "name": "Christian Eitzinger"
    }
  ], 
  "headline": "Probabilistic modelling combined with a CNN for boundary detection of carbon fiber fabrics", 
  "image": "https://zenodo.org/static/img/logos/zenodo-gradient-round.svg", 
  "datePublished": "2019-08-21", 
  "url": "https://zenodo.org/record/3373641", 
  "keywords": [
    "Deep learning, probabilistic modelling, U-Net"
  ], 
  "@context": "https://schema.org/", 
  "identifier": "https://doi.org/10.5281/zenodo.3373641", 
  "@id": "https://doi.org/10.5281/zenodo.3373641", 
  "@type": "ScholarlyArticle", 
  "name": "Probabilistic modelling combined with a CNN for boundary detection of carbon fiber fabrics"
}
34
26
views
downloads
All versions This version
Views 3434
Downloads 2626
Data volume 92.2 MB92.2 MB
Unique views 3333
Unique downloads 2525

Share

Cite as