Explainable Artificial Intelligence for Object Detection in the Automotive Sector
In the automotive domain, object detection is pivotal for enhancing safety and autonomy through the identification of various objects of interest. However, insights into the influential image pixels in the detection process are often lacking. Recognizing these significant regions within the image not only enriches our qualitative understanding of the model’s functionality but also empowers us to refine and optimize its performance. Employing Explainable Artificial Intelligence (XAI), we present an XAI component in this paper. This component explains the predictions made by a pre-trained object detection model for a given image by generating heatmaps that highlight the most critical regions in the image for the detected objects.
Authors
- Panagiotis Radoglou‐Grammatikis (ORCID: https://orcid.org/0000-0003-1605-9413)
- Sotirios K. Goudos (ORCID: https://orcid.org/0000-0001-5981-5683)
- Konstantinos-Filippos Kollias (ORCID: https://orcid.org/0000-0002-3300-5372)
- Panagiotis Sarigiannidis (ORCID: https://orcid.org/0000-0001-6042-0355)
- George F. Fragulis (ORCID: https://orcid.org/0000-0002-8961-7423)
- Θωμάς Λάγκας (ORCID: https://orcid.org/0000-0002-0749-9794)
- Kostas E. Psannis (ORCID: https://orcid.org/0000-0003-0020-6394)
- Marios Siganos (ORCID: https://orcid.org/0000-0001-5318-151X)
- Vasileios Argyriou (ORCID: https://orcid.org/0000-0003-1092-9465)
Institutions
- Democritus University of Thrace (GR)
- University of Macedonia (GR)
- Kingston University London (GB)
- Aristotle University of Thessaloniki (GR)
- Technological Educational Institute of Eastern Macedonia and Thrace (GR)
- Sofia University "St. Kliment Ohridski" (BG)
- University of Western Macedonia (GR)
Publication Details
- Published
- 2025-09-01
- DOI
- https://doi.org/10.3390/engproc2025107044
- Primary Topic
- Advanced Neural Network Applications
- Type
- article
- Field-Weighted Citation Impact
- 0.00