Shortcut Learning in Medical Image Segmentation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lin, Manxi, Weng, Nina, Mikolaj, Kamil, Bashir, Zahra, Svendsen, Morten Bo Søndergaard, Tolsgaard, Martin, Christensen, Anders Nymark, Feragen, Aasa
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914849653522432
author Lin, Manxi
Weng, Nina
Mikolaj, Kamil
Bashir, Zahra
Svendsen, Morten Bo Søndergaard
Tolsgaard, Martin
Christensen, Anders Nymark
Feragen, Aasa
author_facet Lin, Manxi
Weng, Nina
Mikolaj, Kamil
Bashir, Zahra
Svendsen, Morten Bo Søndergaard
Tolsgaard, Martin
Christensen, Anders Nymark
Feragen, Aasa
contents Shortcut learning is a phenomenon where machine learning models prioritize learning simple, potentially misleading cues from data that do not generalize well beyond the training set. While existing research primarily investigates this in the realm of image classification, this study extends the exploration of shortcut learning into medical image segmentation. We demonstrate that clinical annotations such as calipers, and the combination of zero-padded convolutions and center-cropped training sets in the dataset can inadvertently serve as shortcuts, impacting segmentation accuracy. We identify and evaluate the shortcut learning on two different but common medical image segmentation tasks. In addition, we suggest strategies to mitigate the influence of shortcut learning and improve the generalizability of the segmentation models. By uncovering the presence and implications of shortcuts in medical image segmentation, we provide insights and methodologies for evaluating and overcoming this pervasive challenge and call for attention in the community for shortcuts in segmentation. Our code is public at https://github.com/nina-weng/shortcut_skinseg .
format Preprint
id arxiv_https___arxiv_org_abs_2403_06748
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Shortcut Learning in Medical Image Segmentation
Lin, Manxi
Weng, Nina
Mikolaj, Kamil
Bashir, Zahra
Svendsen, Morten Bo Søndergaard
Tolsgaard, Martin
Christensen, Anders Nymark
Feragen, Aasa
Image and Video Processing
Computer Vision and Pattern Recognition
Machine Learning
Shortcut learning is a phenomenon where machine learning models prioritize learning simple, potentially misleading cues from data that do not generalize well beyond the training set. While existing research primarily investigates this in the realm of image classification, this study extends the exploration of shortcut learning into medical image segmentation. We demonstrate that clinical annotations such as calipers, and the combination of zero-padded convolutions and center-cropped training sets in the dataset can inadvertently serve as shortcuts, impacting segmentation accuracy. We identify and evaluate the shortcut learning on two different but common medical image segmentation tasks. In addition, we suggest strategies to mitigate the influence of shortcut learning and improve the generalizability of the segmentation models. By uncovering the presence and implications of shortcuts in medical image segmentation, we provide insights and methodologies for evaluating and overcoming this pervasive challenge and call for attention in the community for shortcuts in segmentation. Our code is public at https://github.com/nina-weng/shortcut_skinseg .
title Shortcut Learning in Medical Image Segmentation
topic Image and Video Processing
Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2403.06748