3D Reconstruction of Leg Bones from X-Ray Images using CNN-based Feature Analysis

Kim, Hangkee; Lee, Ki-Suk; Lee, DongChun; Baek, Nakhoon

doi:10.1109/ictc46691.2019.8939984

Cited by 19 publications

(9 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…27 Nevertheless, in order to increase the amount of training data, random Gaussian filters were applied to the training samples to obtain our osseous reconstruction model. Kim et al 10 did not use deep learning as a reconstruction tool but instead to predict the weights of the statistical model to reconstruct human leg bones. These bone structures do not share the complexity of the internal structures of the knee and hip which were the scope of our method.…”

Section: Discussionmentioning

confidence: 99%

“…The proposed method, however, underperforms in the fibula bone reconstruction because the tibia occludes this bone in the lateral DRR. Furthermore, due to the significant inter‐subject patella shape variability, 10 its statistical approach is complex and demanding, so mispredicted shapes were seldom observed. Finally, although soft tissue contours seem to be well estimated, the differentiation of the soft tissues, which is observable in the CT scan imaging, completely fails to reconstruct with the proposed method.…”

Section: Discussionmentioning

confidence: 99%

“…8 The authors aim to provide patient-specific volumetric CT imaging from 2D images and use patientspecific data with massive data augmentation to train the models. Moreover, Kim et al 10 recently proposed an alternative for 3D reconstruction from x-ray images using CNN-based feature analysis. The idea is to detect the bones' bounding boxes using a CNN and identify feature points and ellipses that will be used to estimate the weights of the corresponding statistical shape model to reconstruct the 3D surface.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Three‐dimensional image volumes from two‐dimensional digitally reconstructed radiographs: A deep learning approach in lower limb CT scans

2021

View full text Add to dashboard Cite

Purpose: Three-dimensional (3D) reconstructions of the human anatomy have been available for surgery planning or diagnostic purposes for a few years now. The different image modalities usually rely on several consecutive two-dimensional (2D) acquisitions in order to reconstruct the 3D volume. Hence, such acquisitions are expensive, time-demanding and often expose the patient to an undesirable amount of radiation. For such reasons, along the most recent years, several studies have been proposed that extrapolate 3D anatomical features from merely 2D exams such as x rays for implant templating in total knee or hip arthroplasties. Method: The presented study shows an adaptation of a deep learning-based convolutional neural network to reconstruct 3D volumes from a mere 2D digitally reconstructed radiograph from one of the most extensive lower limb computed tomography datasets available. This novel approach is based on an encoder-decoder architecture with skip connections and a multidimensional Gaussian filter as data augmentation technique. Results: The results achieved promising values when compared against the ground truth volumes, quantitatively yielding an average of 0.77 AE 0.05 structured similarity index. Conclusions: A novel deep learning-based approach to reconstruct 3D medical image volumes from a single x-ray image was shown in the present study. The network architecture was validated against the original scans presenting SSIM values of 0.77 AE 0.05 and 0.78 AE 0.06, respectively for the knee and the hip crop.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Three‐dimensional image volumes from two‐dimensional digitally reconstructed radiographs: A deep learning approach in lower limb CT scans

2021

View full text Add to dashboard Cite

show abstract

“…Previous approaches [3,15,6,4] use Statistical Shape Models (SSM) or Statistical Shape and Intensity Models (SSIM) for reconstructing bones from X-ray images. However, optimization for the deformable model parameters might be slow and needs a good initialization point to avoid local maxima [20,13].…”

Section: Convolutionalmentioning

confidence: 99%

“…Then, given one or more X-ray images, a 3D reconstruction of the bones is achieved by optimizing the model parameters to maximize the similarity between its rendered versions to the input X-ray images. Kim et al [13] recently used a deep learning approach for detecting landmarks in X-ray images and triangulating them to 3D points. However, their network does not directly outputs a reconstruction of the bones, and the detected 3D landmarks are only used for initialization of the 3D deformable model parameters.…”

Section: Related Workmentioning

confidence: 99%

End-To-End Convolutional Neural Network for 3D Reconstruction of Knee Bones From Bi-Planar X-Ray Images

Kasten¹,

Doktofsky²,

Kovler³

2020

Preprint

View full text Add to dashboard Cite

We present an end-to-end Convolutional Neural Network (CNN) approach for 3D reconstruction of knee bones directly from two bi-planar X-ray images. Clinically, capturing the 3D models of the bones is crucial for surgical planning, implant fitting, and postoperative evaluation. X-ray imaging significantly reduces the exposure of patients to ionizing radiation compared to Computer Tomography (CT) imaging, and is much more common and inexpensive compared to Magnetic Resonance Imaging (MRI) scanners. However, retrieving 3D models from such 2D scans is extremely challenging. In contrast to the common approach of statistically modeling the shape of each bone, our deep network learns the distribution of the bones' shapes directly from the training images. We train our model with both supervised and unsupervised losses using Digitally Reconstructed Radiograph (DRR) images generated from CT scans. To apply our model to X-Ray data, we use style transfer to transform between X-Ray and DRR modalities. As a result, at test time, without further optimization, our solution directly outputs a 3D reconstruction from a pair of bi-planar X-ray images, while preserving geometric constraints. Our results indicate that our deep learning model is very efficient, generalizes well and produces high quality reconstructions.

show abstract

End-To-End Convolutional Neural Network for 3D Reconstruction of Knee Bones from Bi-planar X-Ray Images

Kasten

Doktofsky

Kovler

2020

Lecture Notes in Computer Science

View full text Add to dashboard Cite

3D Reconstruction of Leg Bones from X-Ray Images using CNN-based Feature Analysis

Cited by 19 publications

References 15 publications

Three‐dimensional image volumes from two‐dimensional digitally reconstructed radiographs: A deep learning approach in lower limb CT scans

Three‐dimensional image volumes from two‐dimensional digitally reconstructed radiographs: A deep learning approach in lower limb CT scans

End-To-End Convolutional Neural Network for 3D Reconstruction of Knee Bones From Bi-Planar X-Ray Images

End-To-End Convolutional Neural Network for 3D Reconstruction of Knee Bones from Bi-planar X-Ray Images

Contact Info

Product

Resources

About