Explaining image enhancement black-box methods through a path planning based algorithm

被引:1
|
作者
Cotogni, Marco [1 ]
Cusano, Claudio [1 ]
机构
[1] Univ Pavia, Dept Elect Comp & Biomed Engn, Via Ferrata 1, I-27100 Pavia, Italy
关键词
Explainable AI; Image enhancement; Heuristic search; Path planning;
D O I
10.1007/s11042-023-15648-8
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Nowadays, image-to-image translation methods, are the state of the art for the enhancement of natural images. Even if they usually show high performance in terms of accuracy, they often suffer from several limitations such as the generation of artifacts and the scalability to high resolutions. Moreover, their main drawback is the completely black-box approach that does not allow to provide the final user with any insight about the enhancement processes applied. In this paper we present a path planning algorithm which provides a step-by-step explanation of the output produced by state of the art enhancement methods. This algorithm, called eXIE, uses a variant of A(*) to emulate the enhancement process of another method through the application of an equivalent sequence of enhancing operators. We applied eXIE to explain the output of several state-of-the-art models trained on the Five-K dataset, obtaining sequences of enhancing operators able to produce very similar results in terms of performance and overcoming the huge limitation of poor interpretability of the best performing algorithms.
引用
收藏
页码:8043 / 8062
页数:20
相关论文
共 50 条
  • [1] Explaining image enhancement black-box methods through a path planning based algorithm
    Marco Cotogni
    Claudio Cusano
    Multimedia Tools and Applications, 2024, 83 : 8043 - 8062
  • [2] Automated Image Reduction for Explaining Black-box Classifiers
    Jiang, Mingyue
    Tang, Chengjian
    Zhang, Xiao-Yi
    Zhao, Yangyang
    Ding, Zuohua
    2023 IEEE INTERNATIONAL CONFERENCE ON SOFTWARE ANALYSIS, EVOLUTION AND REENGINEERING, SANER, 2023, : 367 - 378
  • [3] DDImage: an image reduction based approach for automatically explaining black-box classifiers
    Jiang, Mingyue
    Tang, Chengjian
    Zhang, Xiao-Yi
    Zhao, Yangyang
    Ding, Zuohua
    EMPIRICAL SOFTWARE ENGINEERING, 2024, 29 (05)
  • [4] A MODEL DISTILLATION APPROACH FOR EXPLAINING BLACK-BOX MODELS FOR HYPERSPECTRAL IMAGE CLASSIFICATION
    Taskin, Gulsen
    2022 IEEE INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM (IGARSS 2022), 2022, : 3592 - 3595
  • [5] Explaining black-box classifiers: Properties and functions
    Amgoud, Leila
    INTERNATIONAL JOURNAL OF APPROXIMATE REASONING, 2023, 155 : 40 - 65
  • [6] Explaining Black-box Classification Models with Arguments
    Amgoud, Leila
    2021 IEEE 33RD INTERNATIONAL CONFERENCE ON TOOLS WITH ARTIFICIAL INTELLIGENCE (ICTAI 2021), 2021, : 791 - 795
  • [7] Explaining Black-box Android Malware Detection
    Melis, Marco
    Maiorca, Davide
    Biggio, Battista
    Giacinto, Giorgio
    Roli, Fabio
    2018 26TH EUROPEAN SIGNAL PROCESSING CONFERENCE (EUSIPCO), 2018, : 524 - 528
  • [8] Surrogate-based methods for black-box optimization
    Ky Khac Vu
    D'Ambrosio, Claudia
    Hamadi, Youssef
    Liberti, Leo
    INTERNATIONAL TRANSACTIONS IN OPERATIONAL RESEARCH, 2017, 24 (03) : 393 - 424
  • [9] Explaining the black-box smoothly-A counterfactual approach
    Singla, Sumedha
    Eslami, Motahhare
    Pollack, Brian
    Wallace, Stephen
    Batmanghelich, Kayhan
    MEDICAL IMAGE ANALYSIS, 2023, 84
  • [10] Explaining Black-Box Automated Electrocardiogram Classification to Cardiologists
    Oliveira, Derick M.
    Ribeiro, Antonio H.
    Pedrosa, Joao A. O.
    Paixdo, Gabriela M. M.
    Ribeiro, Antonio L.
    Meira Jr, Wagner
    2020 COMPUTING IN CARDIOLOGY, 2020,