TY - GEN
T1 - Evaluation of Post-hoc Interpretability Methods in Breast Cancer Histopathological Image Classification
AU - Waqas, Muhammad
AU - Maul, Tomas
AU - Ahmed, Amr
AU - Liao, Iman Yi
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.
PY - 2024/5/22
Y1 - 2024/5/22
N2 - Methods for post-hoc interpretability are essential for understanding neural network results. Recent years have seen the emergence of numerous post-hoc techniques, but their application to certain tasks, such as histopathological image classification for breast cancer, can produce varied and unpredictable outcomes. Frameworks for quantitative assessment are essential for evaluating each method’s effectiveness. The implementation of post-hoc interpretability methodologies is however hampered by the shortcomings of current frameworks, particularly in high-risk industries. In this study, the performance levels of several common post-hoc interpretability methods are systematically evaluated and compared in the context of histopathological image classification for breast cancer. The study is based on six post-hoc interpretability methods, 3 datasets, and 3 deep neural network models, compared via a RemOve And Retrain (ROAR) approach. The results show that Shapley value sampling obtains the best overall performance in the context of the chosen breast cancer histopathological image datasets.
AB - Methods for post-hoc interpretability are essential for understanding neural network results. Recent years have seen the emergence of numerous post-hoc techniques, but their application to certain tasks, such as histopathological image classification for breast cancer, can produce varied and unpredictable outcomes. Frameworks for quantitative assessment are essential for evaluating each method’s effectiveness. The implementation of post-hoc interpretability methodologies is however hampered by the shortcomings of current frameworks, particularly in high-risk industries. In this study, the performance levels of several common post-hoc interpretability methods are systematically evaluated and compared in the context of histopathological image classification for breast cancer. The study is based on six post-hoc interpretability methods, 3 datasets, and 3 deep neural network models, compared via a RemOve And Retrain (ROAR) approach. The results show that Shapley value sampling obtains the best overall performance in the context of the chosen breast cancer histopathological image datasets.
KW - Breast cancer
KW - Histopathological images
KW - Post-hoc interpretability
KW - ROAR
UR - http://www.scopus.com/inward/record.url?scp=85195140316&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85195140316&partnerID=8YFLogxK
U2 - 10.1007/978-981-97-1417-9_9
DO - 10.1007/978-981-97-1417-9_9
M3 - Conference proceeding (ISBN)
AN - SCOPUS:85195140316
SN - 9789819714162
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 95
EP - 104
BT - Advances in Brain Inspired Cognitive Systems - 13th International Conference, BICS 2023, Proceedings
A2 - Ren, Jinchang
A2 - Hussain, Amir
A2 - Liao, Iman Yi
A2 - Chen, Rongjun
A2 - Huang, Kaizhu
A2 - Zhao, Huimin
A2 - Liu, Xiaoyong
A2 - Ma, Ping
A2 - Maul, Thomas
PB - Springer Science and Business Media Deutschland GmbH
T2 - 13th International Conference on Brain Inspired Cognitive Systems, BICS 2023
Y2 - 5 August 2023 through 6 August 2023
ER -
