The segmentation of chest X-ray (CXR) anatomy is a crucial yet challenging task, often subject to human errors and demanding extensive time investments. While deep learning has shown promise in medical image segmentation, its efficacy is constrained by the necessity for large annotated datasets, which are difficult and costly to generate in the medical domain. Addressing the growing need for models capable of learning from limited data, this study explores the application of Self-Supervised Learning (SSL), specifically adapting the DINO self-supervised learning framework, to leverage unannotated CXR images. Our approach, which does not rely on supervised fine-tuning, is particularly relevant for scenarios where only a small amount of annotated data is available, as is often the case at the onset of continuous learning initiatives. Employing a ConvNeXt-based architecture, our method demonstrates the potential of SSL in enhancing CXR segmentation by utilizing unannotated data, thereby reducing dependency on extensive annotated datasets. In the multilabel segmentation task, we observed increases in the average Dice Similarity Coefficient (DSC) from 0.24±0.19 to 0.56±0.22 for 3 annotated training cases. This research contributes to the evolving landscape of medical imaging analysis, offering insights into the efficient application of SSL for improving segmentation tasks in CXR images and potentially other medical imaging modalities.
Purpose: Deformable registration problems are conventionally posed in a regularized optimization framework, where balance between fidelity and prescribed regularization usually needs to be tuned for each case. Even so, using a single weight to control regularization strength may be insufficient to reflect spatially variant tissue properties and limit registration performance. In this study, we proposed to incorporate a spatially variant deformation prior into image registration framework using a statistical generative model.
Approach: A generator network is trained in an unsupervised setting to maximize the likelihood of observing the moving and fixed image pairs, using an alternating back-propagation approach. The trained model imposes constraints on deformation and serves as an effective low-dimensional deformation parametrization. During registration, optimization is performed over this learned parametrization, eliminating the need for explicit regularization and tuning. The proposed method was tested against SimpleElastix, DIRNet, and Voxelmorph.
Results: Experiments with synthetic images and simulated CTs showed that our method yielded registration errors significantly lower than SimpleElastix and DIRNet. Experiments with cardiac magnetic resonance images showed that the method encouraged physical and physiological feasibility of deformation. Evaluation with left ventricle contours showed that our method achieved a dice of (0.93 ± 0.03) with significant improvement over all SimpleElastix options, DIRNet, and VoxelMorph. Mean average surface distance was on millimeter level, comparable to the best SimpleElastix setting. The average 3D registration time was 12.78 s, faster than 24.70 s in SimpleElastix.
Conclusions: The learned implicit parametrization could be an efficacious alternative to regularized B-spline model, more flexible in admitting spatial heterogeneity.
Concerns on the risks of radiation dose in the cone-beam breast CT (CBBCT) motivated the development of low dose CBBCT (LdCBBCT). Due to the noisy and inadequate data acquisition in LdCBBCT, the conventional analytical Filtered Back Projection (FBP) algorithm tends to result in severe image artifacts and overwhelming noise. Model-based iterative reconstruction methods managed to reduce artifacts and enhance the signal-to-noise ratio but were unable to recover many fine structures and low contrast objects pertinent to diagnosis and treatment. To maintain the strengths of the model-based optimization framework and overcome its limitations in signal recovery, we adapted a CNN-based iterative reconstruction framework, termed Plugand-Play (PnP) proximal gradient descent (PGD) framework, that incorporated state-of-the-art deep-learningbased denoising algorithms with model-based image reconstruction. The PnP-PGD framework is achieved by combining a least-square data fidelity term for data consistency with a non-local regularization for image smoothness, which was solved via PGD. A deep convolutional neural network (DCNN) was plugged in to substitute the proximal operator of the regularization term. The PnP-PGD was evaluated on LdCBBCT scans of a breast phantom and was compared with Filtered Back Projection (FBP), Total Variation (TV), the BlockMatching 3D-transform shrinkage (BM3D), and the DCNN based post-processing method. Compared with FBP, iterative reconstruction, and BM3D, the proposed PnP-PGD substantially reduced image noise and artifacts. Compared with the DCNN based post-processing method, the PnP-PGD improved image contrast-tonoise ratio (CNR). The proposed PnP-PGD takes advantage of both model-based reconstruction and deeplearning-based denoisers, showing improved image quality.
Deformable registration problems are conventionally posed in a regularized optimization framework, where balance between fidelity and prescribed regularization usually needs to be manually tuned for each case. Even so, using a single weight to control regularization strength may be insufficient to reflect spatially variant tissue properties and limit registration performance. In this study, we propose to incorporate a spatially variant deformation prior into image registration framework using a statistical generative model. A generator network is trained in an unsupervised setting to maximize the likelihood of observing the moving and fixed image pairs, using an alternating back-propagation approach. The trained generative model imposes constraints on deformation and serves as an effective low dimensional deformation parametrization. During registration, optimization is performed over this learned parametrization, eliminating the need for explicit regularization and tuning. The proposed method was tested against a B-spline optimization method SimpleElastix, and an end-to-end learning method DIRNet. Experiment with synthetic images shows that our method yielded a registration error of (0.70±0.05) pixels, significantly lower than (0.86±0.12) pixels in SimpleElastix and (0.81±0.06) pixels in DIRNet. Experiment with 2D cardiac MR images demonstrates that the method completed registration with physically and physiologically more feasible deformations and the performance was close to the best of manually tuned results when evaluated with segmentation masks. The average registration time was 1.72 s, faster than 5.63 s in SimpleElastix.
This paper aimed to investigate if deep image features extracted via sparse autoencoder (SAE) could be used to preoperatively predict histologic grade in pancreatic neuroendocrine tumors (pNETs). In this study, a total of 114 patients from two institutions were involved. The deep image features were extracted based on the sparse autoencoder network via a 2000-time iteration. Considering the possible prediction error due to the small patient data size, we performed 10-fold cross-validation. To find the optimal hidden size, we set the size as a range of 6-10. The maximum relevance minimum redundancy (mRMR) features selection algorithm was used to select the most histologic graderelated image features. Then the radiomics signature was generated by using the selected features with Support Vector Machine (SVM), multivariable logistic regression (MLR) and artificial neural networks (ANN) methods. The prediction performance was evaluated using AUC value.
Concerns over the risks of radiation dose from diagnostic CT motivated the utilization of low dose CT (LdCT). However, due to the extremely low X-ray photon statistics in LdCT, the reconstruction problem is ill-posed and noisecontaminated. Conventional Compressed Sensing (CS) methods have been investigated to enhance the signal-to-noise ratio of LdCT at the cost of image resolution and low contrast object visibility. In this work, we adapted a flexible, iterative reconstruction framework, termed Plug-and-Play (PnP) alternating direction method of multipliers (ADMM), that incorporated state-of-the-art denoising algorithms into model-based image reconstruction. The PnP ADMM framework is achieved by combining a least square data fidelity term with a regularization term for image smoothness and was solved through the ADMM. An off-the-shelf image denoiser, the Block-Matching 3D-transform shrinkage (BM3D) filter, is plugged in to substitute an ADMM module. The PnP ADMM was evaluated on low dose scans of ACR 464 phantom and two lung screening data sets and is compared with the Filtered Back Projection (FBP), the Total Variation (TV), the BM3D post-processing method, and the BM3D regularization method. The proposed framework distinguished the line pairs at 9 lp/cm resolution on the ACR phantom and the fissure line in the left lung, resolving the same or better image details than FBP reconstruction of higher dose scans with up to 18 times less dose. Compared with conventional iterative reconstruction methods resulting in comparable image noise, the proposed method is significantly better at recovering image details and improving low contrast conspicuity.
This paper addresses the problem of dynamic magnetic resonance image (DMRI) reconstruction and motion estimation jointly. Because of the inherent anatomical movements in DMRI acquisition, reconstruction of DMRI using motion estimation/compensation (ME/MC) has been explored under the compressed sensing (CS) scheme. In this paper, by embedding the intensity based optical flow (OF) constraint into the traditional CS scheme, we are able to couple the DMRI reconstruction and motion vector estimation. Moreover, the OF constraint is employed in a specific coarse resolution scale in order to reduce the computational complexity. The resulting optimization problem is then solved using a primal-dual algorithm due to its efficiency when dealing with nondifferentiable problems. Experiments on highly accelerated dynamic cardiac MRI with multiple receiver coils validate the performance of the proposed algorithm.
Regularized nonrigid medical image registration algorithms usually estimate the deformation by minimizing a cost function, consisting of a similarity measure and a penalty term that discourages "unreasonable" deformations. Conventional regularization methods enforce homogeneous smoothness properties of the deformation field; less work has been done to incorporate tissue-type-specific elasticity information. Yet ignoring the elasticity differences between tissue types can result in non-physical results, such as bone warping. Bone structures should move rigidly (locally), unlike the more elastic deformation of soft issues. Existing solutions for this problem either treat different regions of an image independently, which requires precise segmentation and incurs boundary issues; or use an empirical spatial varying "filter" to "correct" the deformation field, which requires the knowledge of a stiffness map and departs from the cost-function formulation. We propose a new approach to incorporate tissue rigidity information into the nonrigid registration problem, by developing a space variant regularization function that encourages the local Jacobian of the deformation to be a nearly orthogonal matrix in rigid image regions, while allowing more elastic deformations elsewhere. For the case of X-ray CT data, we use a simple monotonic increasing function of the CT numbers (in HU) as a "rigidity index" since bones typically have the highest CT numbers. Unlike segmentation-based methods, this approach is flexible enough to account for partial volume effects. Results using a B-spline deformation parameterization illustrate that the proposed approach improves registration accuracy in inhale-exhale CT scans with minimal computational penalty.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.