Using Convolutional Neural Network for edge detection in musculoskeletal ultrasound images

Shaima I. Jabbar, Charles R. Day, Nicholas Heinz, Edward K. Chadwick

Research output: Chapter in Book/Report/Conference proceedingConference contribution

13 Citations (Scopus)

Abstract

Fast and accurate segmentation of musculoskeletal ultrasound images is an on-going challenge. Two principal factors make this task difficult: firstly, the presence of speckle noise arising from the interference that accompanies all coherent imaging approaches; secondly, the sometimes subtle interaction between musculoskeletal components that leads to non-uniformity of the image intensity. Our work presents an investigation of the potential of Convolutional Neural Networks (CNNs) to address both of these problems. CNNs are an effective tool that has previously been used in image processing of several biomedical imaging modalities. However, there is little published material addressing the processing of musculoskeletal ultrasound images. In our work we explore the effectiveness of CNNs when trained to act as a pre-segmentation pixel classifier that determines whether a pixel is an edge or non-edge pixel. Our CNNs are trained using two different ground truth interpretations. The first one uses an automatic Canny edge detector to provide the ground truth image; the second ground truth was obtained using the same image marked-up by an expert anatomist. In this initial study the CNNs have been trained using half of the prepared data from one image, using the other half for testing; validation was also carried out using three unseen ultrasound images. CNN performance was assessed using Mathew's Correlation Coefficient, Sensitivity, Specificity and Accuracy. The results show that CNN performance when using expert ground truth image is better than using Canny ground truth image. Our technique is promising and has the potential to speed-up the image processing pipeline using appropriately trained CNNs.

Original languageEnglish
Title of host publication2016 International Joint Conference on Neural Networks, IJCNN 2016
PublisherIEEE Explore
Pages4619-4626
Number of pages8
ISBN (Electronic)9781509006205, 9781509006199
ISBN (Print)9781509006199, 9781509006212
DOIs
Publication statusPublished - 31 Oct 2016
Event2016 International Joint Conference on Neural Networks, IJCNN 2016 - Vancouver, Canada
Duration: 24 Jul 201629 Jul 2016

Publication series

NameProceedings of the International Joint Conference on Neural Networks
Volume2016-October
ISSN (Electronic)2161-4407

Conference

Conference2016 International Joint Conference on Neural Networks, IJCNN 2016
CountryCanada
CityVancouver
Period24/07/1629/07/16

Fingerprint

Edge detection
Ultrasonics
Neural networks
Pixels
Network performance
Image processing
Imaging techniques
Speckle
Classifiers
Pipelines
Detectors
Testing
Processing

Keywords

  • Convolutional Neural Networks
  • Musculoskeletal model
  • Segmentation
  • Ultrasound image

ASJC Scopus subject areas

  • Software
  • Artificial Intelligence

Cite this

Jabbar, S. I., Day, C. R., Heinz, N., & Chadwick, E. K. (2016). Using Convolutional Neural Network for edge detection in musculoskeletal ultrasound images. In 2016 International Joint Conference on Neural Networks, IJCNN 2016 (pp. 4619-4626). [7727805] (Proceedings of the International Joint Conference on Neural Networks; Vol. 2016-October). IEEE Explore. https://doi.org/10.1109/IJCNN.2016.7727805

Using Convolutional Neural Network for edge detection in musculoskeletal ultrasound images. / Jabbar, Shaima I.; Day, Charles R.; Heinz, Nicholas; Chadwick, Edward K.

2016 International Joint Conference on Neural Networks, IJCNN 2016. IEEE Explore, 2016. p. 4619-4626 7727805 (Proceedings of the International Joint Conference on Neural Networks; Vol. 2016-October).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Jabbar, SI, Day, CR, Heinz, N & Chadwick, EK 2016, Using Convolutional Neural Network for edge detection in musculoskeletal ultrasound images. in 2016 International Joint Conference on Neural Networks, IJCNN 2016., 7727805, Proceedings of the International Joint Conference on Neural Networks, vol. 2016-October, IEEE Explore, pp. 4619-4626, 2016 International Joint Conference on Neural Networks, IJCNN 2016, Vancouver, Canada, 24/07/16. https://doi.org/10.1109/IJCNN.2016.7727805
Jabbar SI, Day CR, Heinz N, Chadwick EK. Using Convolutional Neural Network for edge detection in musculoskeletal ultrasound images. In 2016 International Joint Conference on Neural Networks, IJCNN 2016. IEEE Explore. 2016. p. 4619-4626. 7727805. (Proceedings of the International Joint Conference on Neural Networks). https://doi.org/10.1109/IJCNN.2016.7727805
Jabbar, Shaima I. ; Day, Charles R. ; Heinz, Nicholas ; Chadwick, Edward K. / Using Convolutional Neural Network for edge detection in musculoskeletal ultrasound images. 2016 International Joint Conference on Neural Networks, IJCNN 2016. IEEE Explore, 2016. pp. 4619-4626 (Proceedings of the International Joint Conference on Neural Networks).
@inproceedings{437640267f184ebe814f6d1bdd6d818d,
title = "Using Convolutional Neural Network for edge detection in musculoskeletal ultrasound images",
abstract = "Fast and accurate segmentation of musculoskeletal ultrasound images is an on-going challenge. Two principal factors make this task difficult: firstly, the presence of speckle noise arising from the interference that accompanies all coherent imaging approaches; secondly, the sometimes subtle interaction between musculoskeletal components that leads to non-uniformity of the image intensity. Our work presents an investigation of the potential of Convolutional Neural Networks (CNNs) to address both of these problems. CNNs are an effective tool that has previously been used in image processing of several biomedical imaging modalities. However, there is little published material addressing the processing of musculoskeletal ultrasound images. In our work we explore the effectiveness of CNNs when trained to act as a pre-segmentation pixel classifier that determines whether a pixel is an edge or non-edge pixel. Our CNNs are trained using two different ground truth interpretations. The first one uses an automatic Canny edge detector to provide the ground truth image; the second ground truth was obtained using the same image marked-up by an expert anatomist. In this initial study the CNNs have been trained using half of the prepared data from one image, using the other half for testing; validation was also carried out using three unseen ultrasound images. CNN performance was assessed using Mathew's Correlation Coefficient, Sensitivity, Specificity and Accuracy. The results show that CNN performance when using expert ground truth image is better than using Canny ground truth image. Our technique is promising and has the potential to speed-up the image processing pipeline using appropriately trained CNNs.",
keywords = "Convolutional Neural Networks, Musculoskeletal model, Segmentation, Ultrasound image",
author = "Jabbar, {Shaima I.} and Day, {Charles R.} and Nicholas Heinz and Chadwick, {Edward K.}",
year = "2016",
month = "10",
day = "31",
doi = "10.1109/IJCNN.2016.7727805",
language = "English",
isbn = "9781509006199",
series = "Proceedings of the International Joint Conference on Neural Networks",
publisher = "IEEE Explore",
pages = "4619--4626",
booktitle = "2016 International Joint Conference on Neural Networks, IJCNN 2016",

}

TY - GEN

T1 - Using Convolutional Neural Network for edge detection in musculoskeletal ultrasound images

AU - Jabbar, Shaima I.

AU - Day, Charles R.

AU - Heinz, Nicholas

AU - Chadwick, Edward K.

PY - 2016/10/31

Y1 - 2016/10/31

N2 - Fast and accurate segmentation of musculoskeletal ultrasound images is an on-going challenge. Two principal factors make this task difficult: firstly, the presence of speckle noise arising from the interference that accompanies all coherent imaging approaches; secondly, the sometimes subtle interaction between musculoskeletal components that leads to non-uniformity of the image intensity. Our work presents an investigation of the potential of Convolutional Neural Networks (CNNs) to address both of these problems. CNNs are an effective tool that has previously been used in image processing of several biomedical imaging modalities. However, there is little published material addressing the processing of musculoskeletal ultrasound images. In our work we explore the effectiveness of CNNs when trained to act as a pre-segmentation pixel classifier that determines whether a pixel is an edge or non-edge pixel. Our CNNs are trained using two different ground truth interpretations. The first one uses an automatic Canny edge detector to provide the ground truth image; the second ground truth was obtained using the same image marked-up by an expert anatomist. In this initial study the CNNs have been trained using half of the prepared data from one image, using the other half for testing; validation was also carried out using three unseen ultrasound images. CNN performance was assessed using Mathew's Correlation Coefficient, Sensitivity, Specificity and Accuracy. The results show that CNN performance when using expert ground truth image is better than using Canny ground truth image. Our technique is promising and has the potential to speed-up the image processing pipeline using appropriately trained CNNs.

AB - Fast and accurate segmentation of musculoskeletal ultrasound images is an on-going challenge. Two principal factors make this task difficult: firstly, the presence of speckle noise arising from the interference that accompanies all coherent imaging approaches; secondly, the sometimes subtle interaction between musculoskeletal components that leads to non-uniformity of the image intensity. Our work presents an investigation of the potential of Convolutional Neural Networks (CNNs) to address both of these problems. CNNs are an effective tool that has previously been used in image processing of several biomedical imaging modalities. However, there is little published material addressing the processing of musculoskeletal ultrasound images. In our work we explore the effectiveness of CNNs when trained to act as a pre-segmentation pixel classifier that determines whether a pixel is an edge or non-edge pixel. Our CNNs are trained using two different ground truth interpretations. The first one uses an automatic Canny edge detector to provide the ground truth image; the second ground truth was obtained using the same image marked-up by an expert anatomist. In this initial study the CNNs have been trained using half of the prepared data from one image, using the other half for testing; validation was also carried out using three unseen ultrasound images. CNN performance was assessed using Mathew's Correlation Coefficient, Sensitivity, Specificity and Accuracy. The results show that CNN performance when using expert ground truth image is better than using Canny ground truth image. Our technique is promising and has the potential to speed-up the image processing pipeline using appropriately trained CNNs.

KW - Convolutional Neural Networks

KW - Musculoskeletal model

KW - Segmentation

KW - Ultrasound image

UR - http://www.scopus.com/inward/record.url?scp=85007246766&partnerID=8YFLogxK

U2 - 10.1109/IJCNN.2016.7727805

DO - 10.1109/IJCNN.2016.7727805

M3 - Conference contribution

AN - SCOPUS:85007246766

SN - 9781509006199

SN - 9781509006212

T3 - Proceedings of the International Joint Conference on Neural Networks

SP - 4619

EP - 4626

BT - 2016 International Joint Conference on Neural Networks, IJCNN 2016

PB - IEEE Explore

ER -