• English
  • Deutsch
  • Log In
    Password Login
    Research Outputs
    Fundings & Projects
    Researchers
    Institutes
    Statistics
Repository logo
Fraunhofer-Gesellschaft
  1. Home
  2. Fraunhofer-Gesellschaft
  3. Konferenzschrift
  4. Distortion-Based Transparency Detection Using Deep Learning on a Novel Synthetic Image Dataset
 
  • Details
  • Full
Options
2023
Conference Paper
Title

Distortion-Based Transparency Detection Using Deep Learning on a Novel Synthetic Image Dataset

Abstract
Transparency detection is a hard problem, as suggested by animals and humans flying or running into glass. However, humans seem to be able to learn and improve on the task with experience, begging the question, whether computers are able to do so too. Making a computer learn and understand transparency would be beneficial for moving agents, such as robots or autonomous vehicles. Our contributions are threefold: First, we conducted a perception study to obtain insights about human transparency detection methods, when borders of transparent objects are not visible. Second, based on our study insights we created a novel synthetic dataset called DISTOPIA, which focuses on the warping properties of transparent objects, placed in a variety of natural scenes and contains over 140 000 high resolution images. Third, we modified and trained a deep neural network classification model with an attention module to detect transparency through warping. Our results show that a neural network trained on synthetic data depicting only distortion effects can solve the transparency detection problem and surpasses human performance.
Author(s)
Knauthe, Volker
TU Darmstadt, Fachgebiet Graphisch-Interaktive Systeme  
Pöllabauer, Thomas  orcid-logo
Fraunhofer-Institut für Graphische Datenverarbeitung IGD  
Faller, Katharina
TU Darmstadt, Fachgebiet Graphisch-Interaktive Systeme  
Kraus, Maurice
TU Darmstadt, Fachgebiet Graphisch-Interaktive Systeme  
Wirth, Tristan
TU Darmstadt, Fachgebiet Graphisch-Interaktive Systeme  
Buelow, Max von
TU Darmstadt, Fachgebiet Graphisch-Interaktive Systeme  
Kuijper, Arjan  orcid-logo
Fraunhofer-Institut für Graphische Datenverarbeitung IGD  
Fellner, Dieter
Fraunhofer-Institut für Graphische Datenverarbeitung IGD  
Mainwork
Image Analysis. 23rd Scandinavian Conference, SCIA 2023. Proceedings. Part I  
Conference
Scandinavian Conference on Image Analysis 2023  
DOI
10.1007/978-3-031-31435-3_17
Language
English
Fraunhofer-Institut für Graphische Datenverarbeitung IGD  
Keyword(s)
  • Branche: Automotive Industry

  • Branche: Healthcare

  • Branche: Information Technology

  • Research Line: Computer graphics (CG)

  • Research Line: Computer vision (CV)

  • Research Line: Human computer interaction (HCI)

  • Research Line: Machine learning (ML)

  • LTA: Machine intelligence, algorithms, and data structures (incl. semantics)

  • LTA: Generation, capture, processing, and output of images and 3D models

  • Machine learning

  • Deep Learning

  • Robotics

  • Detection

  • Simulation

  • Cookie settings
  • Imprint
  • Privacy policy
  • Api
  • Contact
© 2024