Light-Convolution Dense Selection U-Net (LDS U-Net) for Ultrasound Lateral Bony Feature Segmentation

Sunetra Banerjee,Steven Su,Sai-Ho Ling,Juan Lyu,Timothy Tin-Yan Lee,De Yang,Yongping Zheng,Hung Fat Frank Leung,Zixun Huang

doi:10.3390/app112110180

Abstract

Scoliosis is a widespread medical condition where the spine becomes severely deformed and bends over time. It mostly affects young adults and may have a permanent impact on them. A periodic assessment, using a suitable modality, is necessary for its early detection. Conventionally, the usually employed modalities include X-ray and MRI, which employ ionising radiation and are expensive. Hence, a non-radiating 3D ultrasound imaging technique has been developed as a safe and economic alternative. However, ultrasound produces low-contrast images that are full of speckle noise, and skilled intervention is necessary for their processing. Given the prevalent occurrence of scoliosis and the limitations of scalability of human expert interventions, an automatic, fast, and low-computation assessment technique is being developed for mass scoliosis diagnosis. In this paper, a novel hybridized light-weight convolutional neural network architecture is presented for automatic lateral bony feature identification, which can help to develop a fully-fledged automatic scoliosis detection system. The proposed architecture, Light-convolution Dense Selection U-Net (LDS U-Net), can accurately segment ultrasound spine lateral bony features, from noisy images, thanks to its capabilities of smartly selecting only the useful information and extracting rich deep layer features from the input image. The proposed model is tested using a dataset of 109 spine ultrasound images. The segmentation result of the proposed network is compared with basic U-Net, Attention U-Net, and MultiResUNet using various popular segmentation indices. The results show that LDS U-Net provides a better segmentation performance compared to the other models. Additionally, LDS U-Net requires a smaller number of parameters and less memory, making it suitable for a large-batch screening process of scoliosis without a high computational requirement.

Highlights

The current study shows that the overall occurrence of adolescent idiopathic scoliosis (AIS) is 0.47–5.2% of the total population [2], and is quite prevalent in many regions such as China (5%), Hong Kong (3–4%), and the USA (2%)
Quantitative and qualitative comparisons of the light dense (LD) model, light-dense inception (LDI) model, and light convolution selection (LCS) model are shown together with the LightConvolution Dense Selection U-Net (LDS U-Net) model to evaluate the importance of key features used in the segmentation of spine ultrasound images with variable shapes and sizes of bony features and to assess the overall effectiveness of the proposed network
As ultrasound images are contaminated with noise and are of low contrast, segmentation work is more challenging in this case when compared to other imaging modalities

Summary

Introduction

Scoliosis is a deformation of the spinal cord, usually in an S or C shape with a curvature generally greater than 10◦ , that occurs in the coronal plane between the dorsal and ventral parts. The degree of curvature is stable but in other cases, it increases over time. This ailment primarily starts from adolescence [1]. There are no early signs of idiopathic scoliosis. As the spinal deformation aggravates, patients develop some physical irregularities such as uneven shoulders, inflated curvature of the spine, disproportional alignment of hips, or back pain and discomfort. The early detection of scoliosis is critical to prevent the worsening of the spine over time

Objectives

Results

Discussion

Conclusion