A Comprehensive Review of Explainable Artificial Intelligence (XAI) in Computer Vision
Sensors · 04.07.2025 · 25 (13) · 4166
DOI: 10.3390/s25134166 ↗Kurz gefasst
Übersichtsarbeit in der Zeitschrift Sensors, online erschienen am 04.07.2025. Verfasst von Zhihan Cheng und 4 weiteren Autor:innen. Beteiligte Einrichtung: Department of Mathematics, College of Science, Mathematics and Technology, Wenzhou-Kean University, Wenzhou 325060, China. Hauptthemen laut MEDLINE: Artificial Intelligence. Die Arbeit wurde bisher 114-mal zitiert (Crossref) und nennt 83 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Wenzhou-Kean University.
Abstract (Originalsprache)
Explainable Artificial Intelligence (XAI) is increasingly important in computer vision, aiming to connect complex model outputs with human understanding. This review provides a focused comparative analysis of representative XAI methods in four main categories, attribution-based, activation-based, perturbation-based, and transformer-based approaches, selected from a broader literature landscape. Attribution-based methods like Grad-CAM highlight key input regions using gradients and feature activation. Activation-based methods analyze the responses of internal neurons or feature maps to identify which parts of the input activate specific layers or units, helping to reveal hierarchical feature representations. Perturbation-based techniques, such as RISE, assess feature importance through input modifications without accessing internal model details. Transformer-based methods, which use self-attention, offer global interpretability by tracing information flow across layers. We evaluate these methods using metrics such as faithfulness, localization accuracy, efficiency, and overlap with medical annotations. We also propose a hierarchical taxonomy to classify these methods, reflecting the diversity of XAI techniques. Results show that RISE has the highest faithfulness but is computationally expensive, limiting its use in real-time scenarios. Transformer-based methods perform well in medical imaging, with high IoU scores, though interpreting attention maps requires care. These findings emphasize the need for context-aware evaluation and hybrid XAI methods balancing interpretability and efficiency. The review ends by discussing ethical and practical challenges, stressing the need for standard benchmarks and domain-specific tuning.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Mathematics, College of Science, Mathematics and Technology, Wenzhou-Kean University, Wenzhou 325060, China.
- Department of Mathematics, College of Science, Mathematics and Technology, Kean University, Union, NJ 07083, USA.
- Department of Computer Sciences, College of Science, Mathematics and Technology, Wenzhou-Kean University, Wenzhou 325060, China.
- Department of Computer Sciences, College of Science, Mathematics and Technology, Kean University, Union, NJ 07083, USA.
ORCID-Profile: Zhihan Cheng, Yue Wu, Yule Li, Lingfeng Cai, Baha Ihnaini
Förderung
- Wenzhou-Kean University
Zitieren
APA
Cheng, Z., Wu, Y., Li, Y., Cai, L., & Ihnaini, B. (2025). A Comprehensive Review of Explainable Artificial Intelligence (XAI) in Computer Vision. Sensors, 25(13), 4166. https://doi.org/10.3390/s25134166
Vancouver
Cheng Z, Wu Y, Li Y, Cai L, Ihnaini B. A Comprehensive Review of Explainable Artificial Intelligence (XAI) in Computer Vision. Sensors. 2025;25(13):4166. doi:10.3390/s25134166
BibTeX
@article{Cheng2025,
author = {Cheng, Zhihan and Wu, Yue and Li, Yule and Cai, Lingfeng and Ihnaini, Baha},
title = {A Comprehensive Review of Explainable Artificial Intelligence (XAI) in Computer Vision},
journal = {Sensors},
year = {2025},
volume = {25},
number = {13},
pages = {4166},
doi = {10.3390/s25134166}
}
EUVIMED