On the use of XAI for CNN model interpretation: a remote sensing case study
In this paper, we investigate the use of Explainable Artificial Intelligence (XAI) methods for the interpretation of two Convolutional Neural Network (CNN) classifiers in the field of remote sensing (RS). Specifically, the SegNet and Unet architectures for RS building information extraction and segm...
| Main Authors: | , , , , |
|---|---|
| Format: | Conference or Workshop Item |
| Published: |
IEEE
2022
|
| _version_ | 1825949095638859776 |
|---|---|
| author | Moradi, Loghman Kalantar, Bahareh Zaryabi, Erfan Hasanpour Abdul Halin, Alfian Ueda, Naonori |
| author_facet | Moradi, Loghman Kalantar, Bahareh Zaryabi, Erfan Hasanpour Abdul Halin, Alfian Ueda, Naonori |
| author_sort | Moradi, Loghman |
| collection | UPM |
| description | In this paper, we investigate the use of Explainable Artificial Intelligence (XAI) methods for the interpretation of two Convolutional Neural Network (CNN) classifiers in the field of remote sensing (RS). Specifically, the SegNet and Unet architectures for RS building information extraction and segmentation are evaluated using a comprehensive array of primary- and layer-attributions XAI methods. The attribution methods are quantitatively evaluated using the sensitivity metric. Based on the visualization of the different XAI methods, Deconvolution and GradCAM results in many of the study areas show reliability. Moreover, these methods are able to accurately interpret both Unet's and SegNet's decisions and managed to analyze and reveal the internal mechanisms in both models (confirmed by the low sensitivity scores). Overall, no single method stood out as the best one. |
| first_indexed | 2024-03-06T08:39:18Z |
| format | Conference or Workshop Item |
| id | upm.eprints-37761 |
| institution | Universiti Putra Malaysia |
| last_indexed | 2024-03-06T08:39:18Z |
| publishDate | 2022 |
| publisher | IEEE |
| record_format | dspace |
| spelling | upm.eprints-377612023-10-05T07:09:47Z http://psasir.upm.edu.my/id/eprint/37761/ On the use of XAI for CNN model interpretation: a remote sensing case study Moradi, Loghman Kalantar, Bahareh Zaryabi, Erfan Hasanpour Abdul Halin, Alfian Ueda, Naonori In this paper, we investigate the use of Explainable Artificial Intelligence (XAI) methods for the interpretation of two Convolutional Neural Network (CNN) classifiers in the field of remote sensing (RS). Specifically, the SegNet and Unet architectures for RS building information extraction and segmentation are evaluated using a comprehensive array of primary- and layer-attributions XAI methods. The attribution methods are quantitatively evaluated using the sensitivity metric. Based on the visualization of the different XAI methods, Deconvolution and GradCAM results in many of the study areas show reliability. Moreover, these methods are able to accurately interpret both Unet's and SegNet's decisions and managed to analyze and reveal the internal mechanisms in both models (confirmed by the low sensitivity scores). Overall, no single method stood out as the best one. IEEE 2022 Conference or Workshop Item PeerReviewed Moradi, Loghman and Kalantar, Bahareh and Zaryabi, Erfan Hasanpour and Abdul Halin, Alfian and Ueda, Naonori (2022) On the use of XAI for CNN model interpretation: a remote sensing case study. In: 2022 IEEE Asia-Pacific Conference on Computer Science and Data Engineering (CSDE), 18-20 Dec. 2022, Gold Coast, Australia. . https://ieeexplore.ieee.org/document/10089337 10.1109/CSDE56538.2022.10089337 |
| spellingShingle | Moradi, Loghman Kalantar, Bahareh Zaryabi, Erfan Hasanpour Abdul Halin, Alfian Ueda, Naonori On the use of XAI for CNN model interpretation: a remote sensing case study |
| title | On the use of XAI for CNN model interpretation: a remote sensing case study |
| title_full | On the use of XAI for CNN model interpretation: a remote sensing case study |
| title_fullStr | On the use of XAI for CNN model interpretation: a remote sensing case study |
| title_full_unstemmed | On the use of XAI for CNN model interpretation: a remote sensing case study |
| title_short | On the use of XAI for CNN model interpretation: a remote sensing case study |
| title_sort | on the use of xai for cnn model interpretation a remote sensing case study |
| work_keys_str_mv | AT moradiloghman ontheuseofxaiforcnnmodelinterpretationaremotesensingcasestudy AT kalantarbahareh ontheuseofxaiforcnnmodelinterpretationaremotesensingcasestudy AT zaryabierfanhasanpour ontheuseofxaiforcnnmodelinterpretationaremotesensingcasestudy AT abdulhalinalfian ontheuseofxaiforcnnmodelinterpretationaremotesensingcasestudy AT uedanaonori ontheuseofxaiforcnnmodelinterpretationaremotesensingcasestudy |