Deformable image registration using convolutional neural networks

Eppenhof, Koen A. J.; Lafarge, Maxime W.; Moeskops, Pim; Veta, Mitko; Pluim, Josien P. W.

doi:10.1117/12.2292443

Cited by 30 publications

(8 citation statements)

References 16 publications

(14 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Alternatively, motion estimation can be recast as a datadriven learning task (6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18), which reduces processing times drastically because trained methods can quickly compute centering the images about the center of the left ventricle and cropping the resulting images to the size 80 3 80 3 16. For each subgroup with 30 participants, 20 were randomly chosen for training and the remaining 10 were used for testing, leaving 100 participants for training and 50 participants for testing.…”

mentioning

confidence: 99%

Implementation and Validation of a Three-dimensional Cardiac Motion Estimation Network

Morales¹,

Izquierdo‐Garcia²,

Aganj³

et al. 2019

Radiology: Artificial Intelligence

View full text Add to dashboard Cite

To describe an unsupervised three-dimensional cardiac motion estimation network (CarMEN) for deformable motion estimation from two-dimensional cine MR images. Materials and Methods: A function was implemented using CarMEN, a convolutional neural network that takes two three-dimensional input volumes and outputs a motion field. A smoothness constraint was imposed on the field by regularizing the Frobenius norm of its Jacobian matrix. CarMEN was trained and tested with data from 150 cardiac patients who underwent MRI examinations and was validated on synthetic (n = 100) and pediatric (n = 33) datasets. CarMEN was compared to five state-of-the-art nonrigid body registration methods by using several performance metrics, including Dice similarity coefficient (DSC) and end-point error. Results: On the synthetic dataset, CarMEN achieved a median DSC of 0.85, which was higher than all five methods (minimummaximum median [or MMM], 0.67-0.84; P , .001), and a median end-point error of 1.7, which was lower than (MMM, 2.1-2.7; P , .001) or similar to (MMM, 1.6-1.7; P. .05) all other techniques. On the real datasets, CarMEN achieved a median DSC of 0.73 for Automated Cardiac Diagnosis Challenge data, which was higher than (MMM, 0.33; P , .0001) or similar to (MMM, 0.72-0.75; P. .05) all other methods, and a median DSC of 0.77 for pediatric data, which was higher than (MMM, 0.71-0.76; P , .0001) or similar to (MMM, 0.77-0.78; P. .05) all other methods. All P values were derived from pairwise testing. For all other metrics, CarMEN achieved better accuracy on all datasets than all other techniques except for one, which had the worst motion estimation accuracy. Conclusion: The proposed deep learning-based approach for three-dimensional cardiac motion estimation allowed the derivation of a motion model that balances motion characterization and image registration accuracy and achieved motion estimation accuracy comparable to or better than that of several state-of-the-art image registration algorithms.

show abstract

mentioning

confidence: 99%

Implementation and Validation of a Three-dimensional Cardiac Motion Estimation Network

Morales¹,

Izquierdo‐Garcia²,

Aganj³

et al. 2019

Radiology: Artificial Intelligence

View full text Add to dashboard Cite

show abstract

“…Therefore, studies tested on the same data set were selected for comparison. The study of Eppenhof using convolutional neural network to study the parameters of thin-plate spline transform [26] and the cascade method [27], and the registration of DLIR framework cascading network based on deep learning algorithm were selected as three groups of control studies. In this study, on one hand, the registration results were visualized, and the registration results of the model were qualitatively compared.…”

Section: Evaluation Methodsmentioning

confidence: 99%

A CNN-Based Approach for Lung 3D-CT Registration

Yang

2020

IEEE Access

View full text Add to dashboard Cite

Deep learning techniques have been applied to certain rigid or non-rigid medical image registration due to its potential advantages in meeting the clinical requirements of real-time and accuracy. Based on the deep learning model, this study aims to explore specific network models suitable for lung CT images. The proposed model took unlabeled 3D image pairs as input, and the convolutional neural network (CNN) was utilized and identified as a function with ability of sharing parameters to obtain displacement field. The image pair could be aligned by applying the acquired displacement field to the target image through spatial transformation. The similarity between the aligned image pair combined with the constraints on the displacement field was taken as the objective function to obtain the optimal parameters. Two models with different depths were designed and the consequent registration effects with different optimization methods and convolution kernel sizes were explored. The results proved that the designs with deeper level using Adam optimizer and smaller convolution kernels in obtaining displacement fields had higher accuracy and stronger robustness. The accuracy of the unsupervised model was comparable to state-of-theart methods, while operating orders of magnitude faster. This study proposed a feasible registration method for lung 3D-CT, and its usefulness in aligning CT images has been demonstrated.

show abstract

“…There are also other methods by which artificial DVFs can be generated. For example, Eppenhof et al sampled random numbers from a specified range in a coarse-to-fine grid to generate artificial motions (94,95), and Guo et al used an error-scaling method to generate a training dataset with a target distribution (96). The registration accuracy of all of these methods was either comparable to or better than that of conventional algorithms, in terms of TRE and the Dice score.…”

Section: Applicationsmentioning

confidence: 99%

A review of deep learning-based three-dimensional medical image registration methods

Xiao¹,

Teng²,

Liu³

et al. 2021

Quant Imaging Med Surg

View full text Add to dashboard Cite

Medical image registration is a vital component of many medical procedures, such as imageguided radiotherapy (IGRT), as it allows for more accurate dose-delivery and better management of side effects. Recently, the successful implementation of deep learning (DL) in various fields has prompted many research groups to apply DL to three-dimensional (3D) medical image registration. Several of these efforts have led to promising results. This review summarized the progress made in DL-based 3D image registration over the past 5 years and identify existing challenges and potential avenues for further research.The collected studies were statistically analyzed based on the region of interest (ROI), image modality, supervision method, and registration evaluation metrics. The studies were classified into three categories: deep iterative registration, supervised registration, and unsupervised registration. The studies are thoroughly reviewed and their unique contributions are highlighted. A summary is presented following a review of each category of study, discussing its advantages, challenges, and trends. Finally, the common challenges for all categories are discussed, and potential future research topics are identified.

show abstract

Deformable image registration using convolutional neural networks

Cited by 30 publications

References 16 publications

Implementation and Validation of a Three-dimensional Cardiac Motion Estimation Network

Implementation and Validation of a Three-dimensional Cardiac Motion Estimation Network

A CNN-Based Approach for Lung 3D-CT Registration

A review of deep learning-based three-dimensional medical image registration methods

Contact Info

Product

Resources

About