Technical note: an R package for fitting sparse neural networks with application in animal breeding.
Ontology highlight
ABSTRACT:
SUBMITTER: Wang Y
PROVIDER: S-EPMC6140926 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:PURPOSE:Recently, several attempts were conducted to transfer deep learning to medical image reconstruction. An increasingly number of publications follow the concept of embedding the computed tomography (CT) reconstruction as a known operator into a neural network. However, most of the approaches presented lack an efficient CT reconstruction framework fully integrated into deep learning environments. As a result, many approaches use workarounds for mathematically unambiguously solvable problems. METHODS:PYRO-NN is a generalized framework to embed known operators into the prevalent deep learning framework Tensorflow. The current status includes state-of-the-art parallel-, fan-, and cone-beam projectors, and back-projectors accelerated with CUDA provided as Tensorflow layers. On top, the framework provides a high-level Python API to conduct FBP and iterative reconstruction experiments with data from real CT systems. RESULTS:The framework provides all necessary algorithms and tools to design end-to-end neural network pipelines with integrated CT reconstruction algorithms. The high-level Python API allows a simple use of the layers as known from Tensorflow. All algorithms and tools are referenced to a scientific publication and are compared to existing non-deep learning reconstruction frameworks. To demonstrate the capabilities of the layers, the framework comes with baseline experiments, which are described in the supplementary material. The framework is available as open-source software under the Apache 2.0 licence at https://github.com/csyben/PYRO-NN. CONCLUSIONS:PYRO-NN comes with the prevalent deep learning framework Tensorflow and allows to setup end-to-end trainable neural networks in the medical image reconstruction context. We believe that the framework will be a step toward reproducible research and give the medical physics community a toolkit to elevate medical image reconstruction with new deep learning techniques.
| S-EPMC6899669 | biostudies-literature
Project description:PurposeManual delineation of organs-at-risk (OARs) in radiotherapy is both time-consuming and subjective. Automated and more accurate segmentation is of the utmost importance in clinical application. The purpose of this study is to further improve the segmentation accuracy and efficiency with a novel network named convolutional neural networks (CNN) Cascades.MethodsCNN Cascades was a two-step, coarse-to-fine approach that consisted of a simple region detector (SRD) and a fine segmentation unit (FSU). The SRD first used a relative shallow network to define the region of interest (ROI) where the organ was located, and then, the FSU took the smaller ROI as input and adopted a deep network for fine segmentation. The imaging data (14,651 slices) of 100 head-and-neck patients with segmentations were used for this study. The performance was compared with the state-of-the-art single CNN in terms of accuracy with metrics of Dice similarity coefficient (DSC) and Hausdorff distance (HD) values.ResultsThe proposed CNN Cascades outperformed the single CNN on accuracy for each OAR. Similarly, for the average of all OARs, it was also the best with mean DSC of 0.90 (SRD: 0.86, FSU: 0.87, and U-Net: 0.85) and the mean HD of 3.0 mm (SRD: 4.0, FSU: 3.6, and U-Net: 4.4). Meanwhile, the CNN Cascades reduced the mean segmentation time per patient by 48% (FSU) and 5% (U-Net), respectively.ConclusionsThe proposed two-step network demonstrated superior performance by reducing the input region. This potentially can be an effective segmentation method that provides accurate and consistent delineation with reduced clinician interventions for clinical applications as well as for quality assurance of a multicenter clinical trial.
| S-EPMC6322972 | biostudies-literature
Project description:PurposeVentilation images can be derived from four-dimensional computed tomography (4DCT) by analyzing the change in HU values and deformable vector fields between different respiration phases of computed tomography (CT). As deformable image registration (DIR) is involved, accuracy of 4DCT-derived ventilation image is sensitive to the choice of DIR algorithms. To overcome the uncertainty associated with DIR, we develop a method based on deep convolutional neural network (CNN) to derive ventilation images directly from the 4DCT without explicit image registration.MethodsA total of 82 sets of 4DCT and ventilation images from patients with lung cancer were used in this study. In the proposed CNN architecture, the CT two-channel input data consist of CT at the end of exhale and the end of inhale phases. The first convolutional layer has 32 different kernels of size 5 × 5 × 5, followed by another eight convolutional layers each of which is equipped with an activation layer (ReLU). The loss function is the mean-squared-error (MSE) to measure the intensity difference between the predicted and reference ventilation images.ResultsThe predicted images were comparable to the label images of the test data. The similarity index, correlation coefficient, and Gamma index passing rate averaged over the tenfold cross validation were 0.880 ± 0.035, 0.874 ± 0.024, and 0.806 ± 0.014, respectively.ConclusionsThe results demonstrate that deep CNN can generate ventilation imaging from 4DCT without explicit deformable image registration, reducing the associated uncertainty.
| S-EPMC7098066 | biostudies-literature
Project description:Artificial Neural Networks (ANNs) are bio-inspired models of neural computation that have proven highly effective. Still, ANNs lack a natural notion of time, and neural units in ANNs exchange analog values in a frame-based manner, a computationally and energetically inefficient form of communication. This contrasts sharply with biological neurons that communicate sparingly and efficiently using isomorphic binary spikes. While Spiking Neural Networks (SNNs) can be constructed by replacing the units of an ANN with spiking neurons (Cao et al., 2015; Diehl et al., 2015) to obtain reasonable performance, these SNNs use Poisson spiking mechanisms with exceedingly high firing rates compared to their biological counterparts. Here we show how spiking neurons that employ a form of neural coding can be used to construct SNNs that match high-performance ANNs and match or exceed state-of-the-art in SNNs on important benchmarks, while requiring firing rates compatible with biological findings. For this, we use spike-based coding based on the firing rate limiting adaptation phenomenon observed in biological spiking neurons. This phenomenon can be captured in fast adapting spiking neuron models, for which we derive the effective transfer function. Neural units in ANNs trained with this transfer function can be substituted directly with adaptive spiking neurons, and the resulting Adaptive SNNs (AdSNNs) can carry out competitive classification in deep neural networks without further modifications. Adaptive spike-based coding additionally allows for the dynamic control of neural coding precision: we show empirically how a simple model of arousal in AdSNNs further halves the average required firing rate and this notion naturally extends to other forms of attention as studied in neuroscience. AdSNNs thus hold promise as a novel and sparsely active model for neural computation that naturally fits to temporally continuous and asynchronous applications.
| S-EPMC6332470 | biostudies-literature
Project description:BackgroundSelecting animals for breeding in the optimum way plays an essential role for the management of genetic resources and in selective breeding of livestock species. It requires to compute the optimum genetic contribution of each selection candidate to the next generation. Current software packages for optimum contribution selection (OCS) are not able to handle the main conflicting objectives of animal breeding programs simultaneously, which includes to increase genetic gain, to increase or to maintain genetic diversity, to recover the original genetic background of endangered breeds with historic introgression, and to maintain or increase genetic diversity at native alleles.ResultsThe free R package optiSel offers functions for estimating the above mentioned parameters from pedigree and marker data, and for solving OCS problems. One parameter can be optimized, whereas the remaining ones can be constrained. The results reveal the optimum numbers of offspring of all selection candidates, and can subsequently be used for mate allocation. Different solvers can be used. Solver slsqp was superior when the genetic diversity at native alleles was to be maximized, whereas solvers cccp and cccp2 were superior for all other OCS problems.ConclusionOptimum contribution selection applied to local breeds requires special attention due to the conflicting objectives of their breeding programs. The free R package optiSel is an easy-to-use software taking these conflicting objectives into account.
| S-EPMC6332575 | biostudies-literature
Project description:PurposeRadiation dose delivered to targets located near the upper abdomen or thorax are significantly affected by respiratory motion, necessitating large margins, limiting dose escalation. Surrogate motion management devices, such as the Real-time Position Management (RPM™) system (Varian Medical Systems, Palo Alto, CA), are commonly used to improve normal tissue sparing. Alternative to current solutions, we have developed and evaluated the feasibility of a real-time position management system that leverages the motion data from the onboard hardware of Apple iOS devices to provide patients with visual coaching with the potential to improve the reproducibility of breathing as well as improve patient compliance and reduce treatment delivery time.Methods and materialsThe iOS application, coined the Instant Respiratory Feedback (IRF) system, was developed in Swift (Apple Inc., Cupertino, CA) using the Core-Motion library and implemented on an Apple iPhone® devices. Operation requires an iPhone®, a three-dimensional printed arm, and a radiolucent projector screen system for feedback. Direct comparison between IRF, which leverages sensor fusion data from the iPhone®, and RPM™, an optical-based system, was performed on multiple respiratory motion phantoms and volunteers. The IRF system and RPM™ camera tracking marker were placed on the same location allowing for simultaneous data acquisition. The IRF surrogate measurement of displacement was compared to the signal trace acquired using RPM™ with univariate linear regressions and Bland-Altman analysis.ResultsPeriodic motion shows excellent agreement between both systems, and subject motion shows good agreement during regular and irregular breathing motion. Comparison of IRF and RPM™ show very similar signal traces that were significantly related across all phantoms, including those motion with different amplitude and frequency, and subjects' waveforms (all r > 0.9, P < 0.0001). We demonstrate the feasibility of performing four-dimensional cone beam computed tomography using IRF which provided similar image quality as RPM™ when reconstructing dynamic motion phantom images.ConclusionsFeasibility of an iOS application to provide real-time respiratory motion is demonstrated. This system generated comparable signal traces to a commercially available system and offers an alternative method to monitor respiratory motion.
| S-EPMC7722016 | biostudies-literature
Project description:The point-by-point scanning mechanism of photoacoustic microscopy (PAM) results in low-speed imaging, limiting the application of PAM. In this work, we propose a method to improve the quality of sparse PAM images using convolutional neural networks (CNNs), thereby speeding up image acquisition while maintaining good image quality. The CNN model utilizes attention modules, residual blocks, and perceptual losses to reconstruct the sparse PAM image, which is a mapping from a 1/4 or 1/16 low-sampling sparse PAM image to a latent fully-sampled one. The model is trained and validated mainly on PAM images of leaf veins, showing effective improvements quantitatively and qualitatively. Our model is also tested using in vivo PAM images of blood vessels of mouse ears and eyes. The results suggest that the model can enhance the quality of the sparse PAM image of blood vessels in several aspects, which facilitates fast PAM and its clinical applications.
| S-EPMC7973247 | biostudies-literature
Project description:Neural networks trained by backpropagation have achieved tremendous successes on numerous intelligent tasks. However, naïve gradient-based training and updating methods on memristors impede applications due to intrinsic material properties. Here, we built a 39 nm 1 Gb phase change memory (PCM) memristor array and quantified the unique resistance drift effect. On this basis, spontaneous sparse learning (SSL) scheme that leverages the resistance drift to improve PCM-based memristor network training is developed. During training, SSL regards the drift effect as spontaneous consistency-based distillation process that reinforces the array weights at the high-resistance state continuously unless the gradient-based method switches them to low resistance. Experiments show that the SSL not only helps the convergence of network with better performance and sparsity controllability without additional computation in handwritten digit classification. This work promotes the learning algorithms with the intrinsic properties of memristor devices, opening a new direction for development of neuromorphic computing chips.
| S-EPMC7803975 | biostudies-literature
Project description:Background Endoscopic endonasal surgery (EES) of the skull base often requires extensive bone work in proximity to critical neurovascular structures. Objective To demonstrate the application of an ultrasonic bone curette during EES. Methods Ten patients with skull base lesions underwent EES from September 2011 to April 2012 at the University of Pittsburgh Medical Center. Most of the bone work was done with high-speed drill and rongeurs. The ultrasonic curette was used to remove specific structures. Results All the patients were submitted to fully endoscopic endonasal procedures and had critical bony structures removed with the ultrasonic bone curette. Two patients with degenerative spine diseases underwent odontoid process removal. Five patients with clival and petroclival tumors underwent posterior clinoid removal. Two patients with anterior fossa tumors underwent crista galli removal. One patient underwent unilateral optic nerve decompression. No mechanical or heat injury resulted from the ultrasonic curette. The surrounding neurovascular structures and soft tissue were preserved in all cases. Conclusion In selected EES, the ultrasonic bone curette was successfully used to remove loose pieces of bone in narrow corridors, adjacent to neurovascular structures, and it has advantages to high-speed drills in these specific situations.
| S-EPMC3969437 | biostudies-other
Project description:Conventional image registration utilizing brain voxel information may be erroneous in a neurosurgical setting due to pathology and surgery-related anatomical distortions. We report a novel application of an automated image registration procedure based on skull segmentation for magnetic resonance imaging (MRI) scans acquired before, during and after surgery (i.e., perioperative). The procedure was implemented to assist analysis of intraoperative brain shift in 11 pediatric epilepsy surgery cases, each of whom had up to five consecutive perioperative MRI scans. The procedure consisted of the following steps: (1) Skull segmentation using tissue classification tools. (2) Estimation of rigid body transformation between image pairs using registration driven by the skull segmentation. (3) Composition of transformations to provide transformations between each scan and a common space. The procedure was validated using locations of three types of reference structural landmarks: the skull pin sites, the eye positions, and the scalp skin surface, detected using the peak intensity gradient. The mean target registration error (TRE) scores by skull pin sites and scalp skin rendering were around 1 mm and <1 mm, respectively. Validation by eye position demonstrated >1 mm TRE scores, suggesting it is not a reliable reference landmark in surgical scenarios. Comparable registration accuracy was achieved between opened and closed skull scan pairs and closed and closed skull scan pairs. Our procedure offers a reliable registration framework for processing intrasubject time series perioperative MRI data, with potential of improving intraoperative MRI-based image guidance in neurosurgical practice. Hum Brain Mapp 37:3530-3543, 2016. © 2016 Wiley Periodicals, Inc.
| S-EPMC6867323 | biostudies-literature