scholarly journals CDeep3M - Plug-and-Play cloud based deep learning for image segmentation of light, electron and X-ray microscopy

2018 ◽  
Author(s):  
Matthias G Haberl ◽  
Christopher Churas ◽  
Lucas Tindall ◽  
Daniela Boassa ◽  
Sebastien Phan ◽  
...  
Keyword(s):  
Image Segmentation ◽  
Deep Neural Networks ◽  
Biological Imaging ◽  
Network Architectures ◽  
X Ray ◽  
Biological Images ◽  
3D Microscopy ◽  
Amazon Web Services ◽  
2D And 3D ◽  
Community Access

AbstractAs biological imaging datasets increase in size, deep neural networks are considered vital tools for efficient image segmentation. While a number of different network architectures have been developed for segmenting even the most challenging biological images, community access is still limited by the difficulty of setting up complex computational environments and processing pipelines, and the availability of compute resources. Here, we address these bottlenecks, providing a ready-to-use image segmentation solution for any lab, with a pre-configured, publicly available, cloud-based deep convolutional neural network on Amazon Web Services (AWS). We provide simple instructions for training and applying CDeep3M for segmentation of large and complex 2D and 3D microscopy datasets of diverse biomedical imaging modalities.

scholarly journals Identifying Skeletal Maturity from X-rays using Deep Neural Networks

2021 ◽  
Vol 15 (1) ◽  
pp. 141-148
Author(s):  
Suprava Patnaik ◽  
Sourodip Ghosh ◽  
Richik Ghosh ◽  
Shreya Sahay
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Deep Neural Network ◽  
Deep Neural Networks ◽  
Skeletal Maturity ◽  
Bone Age ◽  
Network Architectures ◽  
X Rays ◽  
X Ray ◽  
Regression Techniques

Skeletal maturity estimation is routinely evaluated by pediatrics and radiologists to assess growth and hormonal disorders. Methods integrated with regression techniques are incompatible with low-resolution digital samples and generate bias, when the evaluation protocols are implemented for feature assessment on coarse X-Ray hand images. This paper proposes a comparative analysis between two deep neural network architectures, with the base models such as Inception-ResNet-V2 and Xception-pre-trained networks. Based on 12,611 hand X-Ray images of RSNA Bone Age database, Inception-ResNet-V2 and Xception models have achieved R-Squared value of 0.935 and 0.942 respectively. Further, in the same order, the MAE accomplished by the two models are 12.583 and 13.299 respectively, when subjected to very few training instances with negligible chances of overfitting.

X-ray Gabor holography

1995 ◽  
Vol 53 ◽  
pp. 612-613
Author(s):  
Steve Lindaas ◽  
Chris Jacobsen ◽  
Alex Kalinovsky ◽  
Malcolm Howells
Keyword(s):  
Surface Relief ◽  
Biological Imaging ◽  
Specimen Preparation ◽  
X Rays ◽  
X Ray ◽  
Water Window ◽  
Biological Objects ◽  
Transmission Imaging ◽  
Atomic Force ◽  
Electron Microscopes

Soft x-ray microscopy offers an approach to transmission imaging of wet, micron-thick biological objects at a resolution superior to that of optical microscopes and with less specimen preparation/manipulation than electron microscopes. Gabor holography has unique characteristics which make it particularly well suited for certain investigations: it requires no prefocussing, it is compatible with flash x-ray sources, and it is able to use the whole footprint of multimode sources. Our method serves to refine this technique in anticipation of the development of suitable flash sources (such as x-ray lasers) and to develop cryo capabilities with which to reduce specimen damage. Our primary emphasis has been on biological imaging so we use x-rays in the water window (between the Oxygen-K and Carbon-K absorption edges) with which we record holograms in vacuum or in air.The hologram is recorded on a high resolution recording medium; our work employs the photoresist poly(methylmethacrylate) (PMMA). Following resist “development” (solvent etching), a surface relief pattern is produced which an atomic force microscope is aptly suited to image.

SAU-Net: A Unified Network for Cell Counting in 2D and 3D Microscopy Images

2021 ◽  
pp. 1-1
Author(s):  
Yue Guo ◽  
Oleh Krupa ◽  
Jason Stein ◽  
Guorong Wu ◽  
Ashok Krishnamurthy
Keyword(s):  
Cell Counting ◽  
3D Microscopy ◽  
2D And 3D ◽  
Microscopy Images

scholarly journals Single-cell conventional pap smear image classification using pre-trained deep neural network architectures

2021 ◽  
Vol 3 (1) ◽  
Author(s):  
Mohammed Aliy Mohammed ◽  
Fetulhak Abdurahman ◽  
Yodit Abebe Ayalew
Keyword(s):  
Neural Network ◽  
Cervical Cancer ◽  
Computer Vision ◽  
Single Cell ◽  
Deep Neural Network ◽  
Deep Neural Networks ◽  
Pap Smear ◽  
Experimental Result ◽  
Network Architectures ◽  
Average Accuracy

Abstract Background Automating cytology-based cervical cancer screening could alleviate the shortage of skilled pathologists in developing countries. Up until now, computer vision experts have attempted numerous semi and fully automated approaches to address the need. Yet, these days, leveraging the astonishing accuracy and reproducibility of deep neural networks has become common among computer vision experts. In this regard, the purpose of this study is to classify single-cell Pap smear (cytology) images using pre-trained deep convolutional neural network (DCNN) image classifiers. We have fine-tuned the top ten pre-trained DCNN image classifiers and evaluated them using five class single-cell Pap smear images from SIPaKMeD dataset. The pre-trained DCNN image classifiers were selected from Keras Applications based on their top 1% accuracy. Results Our experimental result demonstrated that from the selected top-ten pre-trained DCNN image classifiers DenseNet169 outperformed with an average accuracy, precision, recall, and F1-score of 0.990, 0.974, 0.974, and 0.974, respectively. Moreover, it dashed the benchmark accuracy proposed by the creators of the dataset with 3.70%. Conclusions Even though the size of DenseNet169 is small compared to the experimented pre-trained DCNN image classifiers, yet, it is not suitable for mobile or edge devices. Further experimentation with mobile or small-size DCNN image classifiers is required to extend the applicability of the models in real-world demands. In addition, since all experiments used the SIPaKMeD dataset, additional experiments will be needed using new datasets to enhance the generalizability of the models.

scholarly journals X-ray diffraction analysis of matter taking into account the second harmonic in the scattering of powerful ultrashort pulses of an electromagnetic field

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
M. K. Eseev ◽  
A. A. Goshev ◽  
K. A. Makarova ◽  
D. N. Makarov
Keyword(s):  
Electromagnetic Field ◽  
Diffraction Analysis ◽  
Diffraction Pattern ◽  
Second Harmonic ◽  
Ultrashort Pulses ◽  
X Ray Diffraction ◽  
X Ray ◽  
Scattering Spectra ◽  
Technical Possibility ◽  
2D And 3D

AbstractIt is well known that the scattering of ultrashort pulses (USPs) of an electromagnetic field in the X-ray frequency range can be used in diffraction analysis. When such USPs are scattered by various polyatomic objects, a diffraction pattern appears from which the structure of the object can be determined. Today, there is a technical possibility of creating powerful USP sources and the analysis of the scattering spectra of such pulses is a high-precision instrument for studying the structure of matter. As a rule, such scattering occurs at a frequency close to the carrier frequency of the incident USP. In this work, it is shown that for high-power USPs, where the magnetic component of USPs cannot be neglected, scattering at the second harmonic appears. The scattering of USPs by the second harmonic has a characteristic diffraction pattern which can be used to judge the structure of the scattering object; combining the scattering spectra at the first and second harmonics therefore greatly enhances the diffraction analysis of matter. Scattering spectra at the first and second harmonics are shown for various polyatomic objects: examples considered are 2D and 3D materials such as graphene, carbon nanotubes, and hybrid structures consisting of nanotubes. The theory developed in this work can be applied to various multivolume objects and is quite simple for X-ray structural analysis, because it is based on analytical expressions.

Deep neural networks for improving physical accuracy of 2D and 3D multi-mineral segmentation of rock micro-CT images

2021 ◽  
Vol 104 ◽  
pp. 107185 ◽  
Author(s):  
Ying Da Wang ◽  
Mehdi Shabaninejad ◽  
Ryan T. Armstrong ◽  
Peyman Mostaghimi
Keyword(s):  
Neural Networks ◽  
Deep Neural Networks ◽  
Ct Images ◽  
Micro Ct ◽  
2D And 3D

scholarly journals Doppler Tomography in 2D and 3D of the X-ray Binary Cyg X-1 for June 2007

2011 ◽  
Vol 7 (S282) ◽  
pp. 201-202 ◽  
Author(s):  
O. I. Sharova ◽  
M. I. Agafonov ◽  
E. A. Karitskaya ◽  
N. G. Bochkarev ◽  
S. V. Zharikov ◽  
...  
Keyword(s):  
Binary System ◽  
Spectral Data ◽  
Inclination Angle ◽  
Tomographic Reconstruction ◽  
Orbital Plane ◽  
National Observatory ◽  
Doppler Tomography ◽  
X Ray ◽  
2D And 3D ◽  
First Time

AbstractThe 2D and 3D Doppler tomograms of X-ray binary system Cyg X-1 (V1357 Cyg) were reconstructed from spectral data for the line HeII 4686Å obtained with 2-m telescope of the Peak Terskol Observatory (Russia) and 2.1-m telescope of the Mexican National Observatory in June, 2007. Information about gas motions outside the orbital plane, using all of the three velocity components Vx, Vy, Vz, was obtained for the first time. The tomographic reconstruction was carried out for the system inclination angle of 45°. The equal resolution (50 × 50 × 50 km/s) is realized in this case, in the orbital plane (Vx, Vy) and also in the perpendicular direction Vz. The checkout tomograms were realized also for the inclination angle of 40° because of the angle uncertainty. Two versions of the result showed no qualitative discrepancy. Details of the structures revealed by the 3D Doppler tomogram were analyzed.

scholarly journals Adaptive-illumination STED nanoscopy

2017 ◽  
Vol 114 (37) ◽  
pp. 9797-9802 ◽  
Author(s):  
Jörn Heine ◽  
Matthias Reuss ◽  
Benjamin Harke ◽  
Elisa D’Este ◽  
Steffen J. Sahl ◽  
...  
Keyword(s):  
Light Intensity ◽  
Biological Imaging ◽  
The State ◽  
Lower Intensity ◽  
2D And 3D ◽  
Imaging Conditions ◽  
State Difference

The concepts called STED/RESOLFT superresolve features by a light-driven transfer of closely packed molecules between two different states, typically a nonfluorescent “off” state and a fluorescent “on” state at well-defined coordinates on subdiffraction scales. For this, the applied light intensity must be sufficient to guarantee the state difference for molecules spaced at the resolution sought. Relatively high intensities have therefore been applied throughout the imaging to obtain the highest resolutions. At regions where features are far enough apart that molecules could be separated with lower intensity, the excess intensity just adds to photobleaching. Here, we introduce DyMIN (standing for Dynamic Intensity Minimum) scanning, generalizing and expanding on earlier concepts of RESCue and MINFIELD to reduce sample exposure. The principle of DyMIN is that it only uses as much on/off-switching light as needed to image at the desired resolution. Fluorescence can be recorded at those positions where fluorophores are found within a subresolution neighborhood. By tuning the intensity (and thus resolution) during the acquisition of each pixel/voxel, we match the size of this neighborhood to the structures being imaged. DyMIN is shown to lower the dose of STED light on the scanned region up to ∼20-fold under common biological imaging conditions, and >100-fold for sparser 2D and 3D samples. The bleaching reduction can be converted into accordingly brighter images at <30-nm resolution.

Sign in / Sign up

Export Citation Format

Share Document