DEEPBEAS3D: Deep Learning and B-Spline EXPLICIT Active Surfaces

TY - CHAP

T1 - DEEPBEAS3D

T2 - 2023 IEEE International Ultrasonics Symposium, IUS 2023

AU - Williams, Helena

AU - Pedrosa, João

AU - Asad, Muhammad

AU - Cattani, Laura

AU - Vercauteren, Tom

AU - Deprest, Jan

AU - D'Hooge, Jan

N1 - Publisher Copyright: © 2023 IEEE.

PY - 2023

Y1 - 2023

N2 - Deep learning-based automatic segmentation methods have become state-of-the-art. However, they are often not robust enough for direct clinical application, as domain shifts between training and testing data affect their performance. Failure in automatic segmentation can cause sub-optimal results that require correction. To address these problems, we propose a novel 3D extension of an interactive segmentation framework that represents a segmentation from a convolutional neural network (CNN) as a B-spline explicit active surface (BEAS). BEAS ensures segmentations are smooth in 3D space, increasing anatomical plausibility, while allowing the user to precisely edit the 3D surface. We apply this framework to the task of 3D segmentation of the anal sphincter complex (AS) from transperineal ultrasound (TPUS) images, and compare it to the clinical tool used in the pelvic floor disorder clinic (4D View VOCAL, GE Healthcare; Zipf, Austria). Experimental results show that: 1) the proposed framework gives the user explicit control of the surface contour; 2) the perceived workload calculated via the NASA-TLX index was reduced by 30% compared to VOCAL; and 3) it required 70% (170 seconds) less user time than VOCAL (p< 0.00001).

AB - Deep learning-based automatic segmentation methods have become state-of-the-art. However, they are often not robust enough for direct clinical application, as domain shifts between training and testing data affect their performance. Failure in automatic segmentation can cause sub-optimal results that require correction. To address these problems, we propose a novel 3D extension of an interactive segmentation framework that represents a segmentation from a convolutional neural network (CNN) as a B-spline explicit active surface (BEAS). BEAS ensures segmentations are smooth in 3D space, increasing anatomical plausibility, while allowing the user to precisely edit the 3D surface. We apply this framework to the task of 3D segmentation of the anal sphincter complex (AS) from transperineal ultrasound (TPUS) images, and compare it to the clinical tool used in the pelvic floor disorder clinic (4D View VOCAL, GE Healthcare; Zipf, Austria). Experimental results show that: 1) the proposed framework gives the user explicit control of the surface contour; 2) the perceived workload calculated via the NASA-TLX index was reduced by 30% compared to VOCAL; and 3) it required 70% (170 seconds) less user time than VOCAL (p< 0.00001).

KW - Active Model

KW - B-spline explicit active surfaces

KW - CNN

KW - Interactive Segmentation

KW - Transperineal Ultrasound

UR - http://www.scopus.com/inward/record.url?scp=85178636222&partnerID=8YFLogxK

U2 - 10.1109/IUS51837.2023.10308101

DO - 10.1109/IUS51837.2023.10308101

M3 - Conference paper

AN - SCOPUS:85178636222

T3 - IEEE International Ultrasonics Symposium, IUS

BT - IUS 2023 - IEEE International Ultrasonics Symposium, Proceedings

PB - IEEE Computer Society

Y2 - 3 September 2023 through 8 September 2023

ER -