Automatic segmentation of the femur and tibia bones from X-ray images based on pure dilated residual U-Net

Shen, Weihao; Xu, Wenbo; Zhang, Hongyang; Sun, Zhang‐Hua; Ma, Jianxiong; Ma, Xinlong; Zhou, S. Kevin; Guo, Shuxiang; Wang, Yuanquan

doi:10.3934/ipi.2020057

Cited by 37 publications

(16 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This U-net is a CNN that was developed for biomedical image segmentation [20,21]. The network has demonstrated high performance on computer-aided diagnosis in several prior studies related to bones, and it reduces more outliers because it is not a standardized process [22][23][24]. The created masks including only rib regions by these two methods were examined to label the 12 pairs of ribs using 3D region growing.…”

Section: Methodsmentioning

confidence: 99%

The Development of an Automatic Rib Sequence Labeling System on Axial Computed Tomography Images with 3-Dimensional Region Growing

Seol

Kim

Park

et al. 2022

Sensors

View full text Add to dashboard Cite

This paper proposes a development of automatic rib sequence labeling systems on chest computed tomography (CT) images with two suggested methods and three-dimensional (3D) region growing. In clinical practice, radiologists usually define anatomical terms of location depending on the rib’s number. Thus, with the manual process of labeling 12 pairs of ribs and counting their sequence, it is necessary to refer to the annotations every time the radiologists read chest CT. However, the process is tedious, repetitive, and time-consuming as the demand for chest CT-based medical readings has increased. To handle the task efficiently, we proposed an automatic rib sequence labeling system and implemented comparison analysis on two methods. With 50 collected chest CT images, we implemented intensity-based image processing (IIP) and a convolutional neural network (CNN) for rib segmentation on this system. Additionally, three-dimensional (3D) region growing was used to classify each rib’s label and put in a sequence label. The IIP-based method reported a 92.0% and the CNN-based method reported a 98.0% success rate, which is the rate of labeling appropriate rib sequences over whole pairs (1st to 12th) for all slices. We hope for the applicability thereof in clinical diagnostic environments by this method-efficient automatic rib sequence labeling system.

show abstract

Section: Methodsmentioning

confidence: 99%

The Development of an Automatic Rib Sequence Labeling System on Axial Computed Tomography Images with 3-Dimensional Region Growing

Seol

Kim

Park

et al. 2022

Sensors

View full text Add to dashboard Cite

show abstract

“…Traditionally, segmentation of 2-D radiographic images has been beneficial but rarely studied owing to its difficulties, including poor and non-uniform image contrast, noise, occlusions, and overlap of neighboring structures. Recently, deep learning-based methods widely used for medical image segmentation have been applied to hip joint segmentation, with excellent results [47], [48]. Shen et al proposed a pure dilated residual U-net that improved the accuracy and convergence of vanilla U-net [21] in segmenting femurs and tibias [47].…”

Section: Related Researchmentioning

confidence: 99%

Deep Learning and Computer Vision Techniques for Automated Total Hip Arthroplasty Planning on 2-D Radiographs

Kim

Yoon

2022

IEEE Access

View full text Add to dashboard Cite

Preoperative planning is mandatory for successful total hip arthroplasty (THA). In planning, the operating surgeon should decide the best type and size of THA components for the patient. However, most digital templating software only simulates acetate templating by overlaying the shape of the prosthesis components on a radiograph; the selection and positioning of the prostheses are performed manually depending on the operator's experience. Determining the optimal type and size of THA components is a repetitive and time-consuming task for digital and acetate templating. This study proposes a novel approach to automatically select and position THA components that are most suitable for the patient's bone anatomy. The approach consists of two phases: segmenting a hip anteroposterior (AP) radiographic image into five predefined anatomical regions using a fully convolutional neural network, and estimating the optimal sizes and positions of THA components using deep learning and computer vision technology. The experiments demonstrated that the accuracy of acetabular and femoral component size prediction within one size error was 78.9% and 70.9%, respectively. Compared with meta-analysis results from previous studies, our results are close to human level. An automated digital templating prototype system was developed using our research results and tested in a clinical setting to evaluate field adaptability. These processes are introduced in this study.INDEX TERMS Artificial intelligence (AI), automated 2-D templating, hip joint segmentation, implant size estimation, prostheses recommendation, total hip arthroplasty (THA)

show abstract

“…The feature maps learned by this layer is directly passed to all subsequent layers as input, which alleviates the gradient disappearance or explosion caused by the transmission of input information and gradient information among multiple layers, and promotes the formation of deeper networks. Shen et al(2020) [43] proposed a new segmentation approach based on pure dilated residual U-Net. The dilated convolution is completely employed to increase the receptive field.…”

Section: Related Workmentioning

confidence: 99%

Segmentation of Biventricle in Cardiac Cine MRI via Nested Capsule Dense Network

Hu¹,

Zhang²,

Zhang³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Cardiac magnetic resonance image (MRI) has been widely used in diagnosis of cardiovascular diseases because of its noninvasive nature and high image quality. The evaluation standard of physiological indexes in cardiac diagnosis is essentially the accuracy of segmentation of left ventricle (LV) and right ventricle (RV) in cardiac MRI. In this paper, we propose a novel Nested Capsule Dense Network (NCDN) structure based on the FC-DenseNet model and capsule convolution-capsule deconvolution. Different from the traditional symmetric single codec network structure such as U-net, NCDN uses multiple codecs instead of a single codec to achieve multi-resolution, which makes it possible to save more spatial information and improve the robustness of the model. The proposed model is tested on three datasets that includes York University Cardiac MRI dataset, Automated Cardiac Diagnosis Challenge (ACDC-2017), and local dataset. The results show that the proposed NCDN outperforms the state-of-the-art methods.

show abstract

Automatic segmentation of the femur and tibia bones from X-ray images based on pure dilated residual U-Net

Cited by 37 publications

References 33 publications

The Development of an Automatic Rib Sequence Labeling System on Axial Computed Tomography Images with 3-Dimensional Region Growing

The Development of an Automatic Rib Sequence Labeling System on Axial Computed Tomography Images with 3-Dimensional Region Growing

Deep Learning and Computer Vision Techniques for Automated Total Hip Arthroplasty Planning on 2-D Radiographs

Segmentation of Biventricle in Cardiac Cine MRI via Nested Capsule Dense Network

Contact Info

Product

Resources

About