“…Especially, considering 3D reconstruction often requires comparatively long computation times. As it stands, memory and computation requirements for training 3D neural networks can be orders of magnitude larger than their 2D counterparts 57 . As such, challenges regarding network size, as well as stability during training, are major hurdles for successful implementation.…”
Section: Compressed Sensingmentioning
confidence: 99%
“…As it stands, memory and computation requirements for training 3D neural networks can be orders of magnitude larger than their 2D counterparts. 57 As such, challenges regarding network size, as well as stability during training, are major hurdles for successful implementation. With these constraints in mind, Pham et al 52 adapt a 2D neural network (Super-Resolution Convolutional Neural Network (SRCNN) 58 ) into a 3D equivalent (SRCNN3D) to perform 3D image reconstruction.…”
Summary
Magnetic resonance (MR) imaging visualises soft tissue contrast in exquisite detail without harmful ionising radiation. In this work, we provide a state‐of‐the‐art review on the use of deep learning in MR image reconstruction from different image acquisition types involving compressed sensing techniques, parallel image acquisition and multi‐contrast imaging. Publications with deep learning‐based image reconstruction for MR imaging were identified from the literature (PubMed and Google Scholar), and a comprehensive description of each of the works was provided. A detailed comparison that highlights the differences, the data used and the performance of each of these works were also made. A discussion of the potential use cases for each of these methods is provided. The sparse image reconstruction methods were found to be most popular in using deep learning for improved performance, accelerating acquisitions by around 4–8 times. Multi‐contrast image reconstruction methods rely on at least one pre‐acquired image, but can achieve 16‐fold, and even up to 32‐ to 50‐fold acceleration depending on the set‐up. Parallel imaging provides frameworks to be integrated in many of these methods for additional speed‐up potential. The successful use of compressed sensing techniques and multi‐contrast imaging with deep learning and parallel acquisition methods could yield significant MR acquisition speed‐ups within clinical routines in the near future.
“…Especially, considering 3D reconstruction often requires comparatively long computation times. As it stands, memory and computation requirements for training 3D neural networks can be orders of magnitude larger than their 2D counterparts 57 . As such, challenges regarding network size, as well as stability during training, are major hurdles for successful implementation.…”
Section: Compressed Sensingmentioning
confidence: 99%
“…As it stands, memory and computation requirements for training 3D neural networks can be orders of magnitude larger than their 2D counterparts. 57 As such, challenges regarding network size, as well as stability during training, are major hurdles for successful implementation. With these constraints in mind, Pham et al 52 adapt a 2D neural network (Super-Resolution Convolutional Neural Network (SRCNN) 58 ) into a 3D equivalent (SRCNN3D) to perform 3D image reconstruction.…”
Summary
Magnetic resonance (MR) imaging visualises soft tissue contrast in exquisite detail without harmful ionising radiation. In this work, we provide a state‐of‐the‐art review on the use of deep learning in MR image reconstruction from different image acquisition types involving compressed sensing techniques, parallel image acquisition and multi‐contrast imaging. Publications with deep learning‐based image reconstruction for MR imaging were identified from the literature (PubMed and Google Scholar), and a comprehensive description of each of the works was provided. A detailed comparison that highlights the differences, the data used and the performance of each of these works were also made. A discussion of the potential use cases for each of these methods is provided. The sparse image reconstruction methods were found to be most popular in using deep learning for improved performance, accelerating acquisitions by around 4–8 times. Multi‐contrast image reconstruction methods rely on at least one pre‐acquired image, but can achieve 16‐fold, and even up to 32‐ to 50‐fold acceleration depending on the set‐up. Parallel imaging provides frameworks to be integrated in many of these methods for additional speed‐up potential. The successful use of compressed sensing techniques and multi‐contrast imaging with deep learning and parallel acquisition methods could yield significant MR acquisition speed‐ups within clinical routines in the near future.
“…Livermore Big Artificial Neural Network can spatially partition the training over many graphics processing unit (GPU)–accelerated HPC nodes, enabling the traditional robust scaling that other HPC applications enjoy, that is, accelerated time to solution without a compromise in the quality of the learned model (Van Essen et al, 2015). For the CosmoFlow problem, LBANN is able to achieve an order-of-magnitude improvement in prediction quality using the full 3D data sets in training while significantly reducing training time by exploiting a much larger-scale system (Oyama et al, 2020). The GAN-based surrogate models should be able to take advantage of LBANN to an even greater degree.Figure 3.Logarithmic histograms of pixel intensities from the recurrent neural networks–generated and validation cosmology data sets show an excellent match.…”
Rapid growth in data, computational methods, and computing power is driving a remarkable revolution in what variously is termed machine learning (ML), statistical learning, computational learning, and artificial intelligence. In addition to highly visible successes in machine-based natural language translation, playing the game Go, and self-driving cars, these new technologies also have profound implications for computational and experimental science and engineering, as well as for the exascale computing systems that the Department of Energy (DOE) is developing to support those disciplines. Not only do these learning technologies open up exciting opportunities for scientific discovery on exascale systems, they also appear poised to have important implications for the design and use of exascale computers themselves, including high-performance computing (HPC) for ML and ML for HPC. The overarching goal of the ExaLearn co-design project is to provide exascale ML software for use by Exascale Computing Project (ECP) applications, other ECP co-design centers, and DOE experimental facilities and leadership class computing facilities.
“…Data parallelism has been also used for COVID-19 diagnosis based on CT scans [14], text and feature extraction based diagnosis using CNN models [15], [16]. Some studies combine some of the abovementioned techniques, called hybrid parallelism, to handle 3D images and models [17], [18]. In [19] a scalable toolkit for medical image segmentation is presented, but is privative and only two models are provided.…”
Most research on novel techniques for 3D Medical Image Segmentation (MIS) is currently done using Deep Learning with GPU accelerators. The principal challenge of such technique is that a single input can easily cope computing resources, and require prohibitive amounts of time to be processed. Distribution of deep learning and scalability over computing devices is an actual need for progressing on such research field. Conventional distribution of neural networks consist in "data parallelism", where data is scattered over resources (e.g., GPUs) to parallelize the training of the model. However, "experiment parallelism" is also an option, where different training processes (i.e., on a hyper-parameter search) are parallelized across resources. While the first option is much more common on 3D image segmentation, the second provides a pipeline design with less dependence among parallelized processes, allowing overhead reduction and more potential scalability. In this work we present a design for distributed deep learning training pipelines, focusing on multinode and multi-GPU environments, where the two different distribution approaches are deployed and benchmarked. We take as proof of concept the 3D U-Net architecture, using the MSD Brain Tumor Segmentation dataset, a state-of-art problem in medical image segmentation with high computing and space requirements. Using the BSC MareNostrum supercomputer as benchmarking environment, we use TensorFlow and Ray as neural network training and experiment distribution platforms. We evaluate the experiment speed-up when parallelizing, showing the potential for scaling out on GPUs and nodes. Also comparing the different parallelism techniques, showing how experiment distribution leverages better such resources through scaling, e.g. by a speed-up factor from x12 to x14 using 32 GPUs. Finally, we provide the implementation of the design open to the community, and the non-trivial steps and methodology for adapting and deploying a MIS case as the here presented.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
