TY - JOUR
T1 - Ultrasound spine image segmentation using multi-scale feature fusion Skip-Inception U-Net (SIU-Net)
AU - Banerjee, Sunetra
AU - Lyu, Juan
AU - Huang, Zixun
AU - Leung, Frank H.F.
AU - Lee, Timothy
AU - Yang, De
AU - Su, Steven
AU - Zheng, Yongping
AU - Ling, Sai Ho
N1 - Funding Information:
The project is partially supported by Hong Kong Research Grant Council, Research Impact Fund (R5017-18).
Publisher Copyright:
© 2022 Nalecz Institute of Biocybernetics and Biomedical Engineering of the Polish Academy of Sciences
PY - 2022/1/1
Y1 - 2022/1/1
N2 - Scoliosis is a 3D spinal deformation where the spine takes a lateral curvature, forming an angle in the coronal plane. Diagnosis of scoliosis requires periodic detection, and frequent exposure to radiative imaging may cause cancer. A safer and more economical alternative imaging, i.e., 3D ultrasound imaging modality, is being explored. However, unlike other radiative modalities, an ultrasound image is noisy, which often suppresses the image's useful information. Through this research, a novel hybridized CNN architecture, multi-scale feature fusion Skip-Inception U-Net (SIU-Net), is proposed for a fully automatic bony feature detection, which can be further used to assess the severity of scoliosis safely and automatically. The proposed architecture, SIU-Net, incorporates two novel features into the basic U-Net architecture: (a) an improvised Inception block and (b) newly designed decoder-side dense skip pathways. The proposed model is tested on 109 spine ultrasound image datasets. The architecture is evaluated using the popular (i) Jaccard Index (ii) Dice Coefficient and (iii) Euclidean distance, and compared with (a) the basic U-net segmentation model, (b) a more evolved UNet++ model, and (c) a newly developed MultiResUNet model. The results show that SIU-Net gives the clearest segmentation output, especially in the important regions of interest such as thoracic and lumbar bony features. The method also gives the highest average Jaccard score of 0.781 and Dice score of 0.883 and the lowest histogram Euclidean distance of 0.011 than the other three models. SIU-Net looks promising to meet the objectives of a fully automatic scoliosis detection system.
AB - Scoliosis is a 3D spinal deformation where the spine takes a lateral curvature, forming an angle in the coronal plane. Diagnosis of scoliosis requires periodic detection, and frequent exposure to radiative imaging may cause cancer. A safer and more economical alternative imaging, i.e., 3D ultrasound imaging modality, is being explored. However, unlike other radiative modalities, an ultrasound image is noisy, which often suppresses the image's useful information. Through this research, a novel hybridized CNN architecture, multi-scale feature fusion Skip-Inception U-Net (SIU-Net), is proposed for a fully automatic bony feature detection, which can be further used to assess the severity of scoliosis safely and automatically. The proposed architecture, SIU-Net, incorporates two novel features into the basic U-Net architecture: (a) an improvised Inception block and (b) newly designed decoder-side dense skip pathways. The proposed model is tested on 109 spine ultrasound image datasets. The architecture is evaluated using the popular (i) Jaccard Index (ii) Dice Coefficient and (iii) Euclidean distance, and compared with (a) the basic U-net segmentation model, (b) a more evolved UNet++ model, and (c) a newly developed MultiResUNet model. The results show that SIU-Net gives the clearest segmentation output, especially in the important regions of interest such as thoracic and lumbar bony features. The method also gives the highest average Jaccard score of 0.781 and Dice score of 0.883 and the lowest histogram Euclidean distance of 0.011 than the other three models. SIU-Net looks promising to meet the objectives of a fully automatic scoliosis detection system.
KW - Bony feature
KW - Convolutional neural network
KW - Feature fusion
KW - Scoliosis
KW - Segmentation
KW - U-Net
KW - Ultrasound
UR - http://www.scopus.com/inward/record.url?scp=85126028727&partnerID=8YFLogxK
U2 - 10.1016/j.bbe.2022.02.011
DO - 10.1016/j.bbe.2022.02.011
M3 - Journal article
AN - SCOPUS:85126028727
SN - 0208-5216
VL - 42
SP - 341
EP - 361
JO - Biocybernetics and Biomedical Engineering
JF - Biocybernetics and Biomedical Engineering
IS - 1
ER -