Quality control for more reliable integration of deep learning-based image 
segmentation into medical workflows

Williams, Elena; Niehaus, Sebastian; Reinelt, Janis; Merola, Alberto; Mihai, Paul Glad; Roeder, Ingo; Scherf, Nico; Hernández, Maria del C. Valdés

doi:10.48550/arXiv.2112.03277

アイテム詳細

登録内容を編集ファイル形式で保存

一時保存へ追加

タグ情報を表示リリース履歴を表示詳細要約

公開

Preprint

Quality control for more reliable integration of deep learning-based image segmentation into medical workflows

MPS-Authors

There are no MPG-Authors in the publication available

External Resource

https://arxiv.org/abs/2112.03277v2
(プレプリント)

Fulltext (restricted access)

There are currently no full texts shared for your IP range.

フルテキスト (公開)

公開されているフルテキストはありません

付随資料 (公開)

There is no public supplementary material available

引用

Williams, E., Niehaus, S., Reinelt, J., Merola, A., Mihai, P. G., Roeder, I., Scherf, N., & Hernández, M. d. C. V. (2021). Quality control for more reliable integration of deep learning-based image segmentation into medical workflows. arXiv. doi:10.48550/arXiv.2112.03277.

引用: https://hdl.handle.net/21.11116/0000-000A-E97E-3

要旨

Machine learning algorithms underpin modern diagnostic-aiding software, which
has proved valuable in clinical practice, particularly in radiology. However,
inaccuracies, mainly due to the limited availability of clinical samples for
training these algorithms, hamper their wider applicability, acceptance, and
recognition amongst clinicians. We present an analysis of state-of-the-art
automatic quality control (QC) approaches that can be implemented within these
algorithms to estimate the certainty of their outputs. We validated the most
promising approaches on a brain image segmentation task identifying white
matter hyperintensities (WMH) in magnetic resonance imaging data. WMH are a
correlate of small vessel disease common in mid-to-late adulthood and are
particularly challenging to segment due to their varied size, and
distributional patterns. Our results show that the aggregation of uncertainty
and Dice prediction were most effective in failure detection for this task.
Both methods independently improved mean Dice from 0.82 to 0.84. Our work
reveals how QC methods can help to detect failed segmentation cases and
therefore make automatic segmentation more reliable and suitable for clinical
practice.