An iterative reconstruction method for sparse-projection data for low-dose CT

2021 ◽  
pp. 1-16
Author(s):  
Ying Huang ◽  
Qian Wan ◽  
Zixiang Chen ◽  
Zhanli Hu ◽  
Guanxun Cheng ◽  
...  
Keyword(s):  
Signal To Noise Ratio ◽  
Structural Similarity ◽  
Reconstructed Image ◽  
Projection Data ◽  
Reconstruction Method ◽  
X Ray ◽  
Image Structure ◽  
Reconstruction Performance ◽  
Risk Of Cancer ◽  
Suppress Noise

Reducing X-ray radiation is beneficial for reducing the risk of cancer in patients. There are two main approaches for achieving this goal namely, one is to reduce the X-ray current, and another is to apply sparse-view protocols to do image scanning and projections. However, these techniques usually lead to degradation of the reconstructed image quality, resulting in excessive noise and severe edge artifacts, which seriously affect the diagnosis result. In order to overcome such limitation, this study proposes and tests an algorithm based on guided kernel filtering. The algorithm combines the characteristics of anisotropic edges between adjacent image voxels, expresses the relevant weights with an exponential function, and adjusts the weights adaptively through local gray gradients to better preserve the image structure while suppressing noise information. Experiments show that the proposed method can effectively suppress noise and preserve the image structure. Comparing with similar algorithms, the proposed algorithm greatly improves the peak signal-to-noise ratio (PSNR), structural similarity (SSIM), and root mean square error (RMSE) of the reconstructed image. The proposed algorithm has the best effect in quantitative analysis, which verifies the effectiveness of the proposed method and good image reconstruction performance. Overall, this study demonstrates that the proposed method can reduce the number of projections required for repeated CT scans and has potential for medical applications in reducing radiation doses.

scholarly journals Encryption Based Image Watermarking Algorithm in 2DWT-DCT Domains

Sensors ◽  
2021 ◽  
Vol 21 (16) ◽  
pp. 5540
Author(s):  
Nayeem Hasan ◽  
Md Saiful Islam ◽  
Wenyu Chen ◽  
Muhammad Ashad Kabir ◽  
Saad Al-Ahmadi
Keyword(s):  
Signal To Noise Ratio ◽  
Image Watermarking ◽  
Similarity Index ◽  
Structural Similarity ◽  
Reconstructed Image ◽  
Discrete Wavelet ◽  
Signal To Noise ◽  
Noise Ratio ◽  
Watermarking Scheme

This paper proposes an encryption-based image watermarking scheme using a combination of second-level discrete wavelet transform (2DWT) and discrete cosine transform (DCT) with an auto extraction feature. The 2DWT has been selected based on the analysis of the trade-off between imperceptibility of the watermark and embedding capacity at various levels of decomposition. DCT operation is applied to the selected area to gather the image coefficients into a single vector using a zig-zig operation. We have utilized the same random bit sequence as the watermark and seed for the embedding zone coefficient. The quality of the reconstructed image was measured according to bit correction rate, peak signal-to-noise ratio (PSNR), and similarity index. Experimental results demonstrated that the proposed scheme is highly robust under different types of image-processing attacks. Several image attacks, e.g., JPEG compression, filtering, noise addition, cropping, sharpening, and bit-plane removal, were examined on watermarked images, and the results of our proposed method outstripped existing methods, especially in terms of the bit correction ratio (100%), which is a measure of bit restoration. The results were also highly satisfactory in terms of the quality of the reconstructed image, which demonstrated high imperceptibility in terms of peak signal-to-noise ratio (PSNR ≥ 40 dB) and structural similarity (SSIM ≥ 0.9) under different image attacks.

scholarly journals SARA-GAN: Self-Attention and Relative Average Discriminator Based Generative Adversarial Networks for Fast Compressed Sensing MRI Reconstruction

2020 ◽  
Vol 14 ◽  
Author(s):  
Zhenmou Yuan ◽  
Mingfeng Jiang ◽  
Yaming Wang ◽  
Bo Wei ◽  
Yongming Li ◽  
...  
Keyword(s):  
Signal To Noise Ratio ◽  
Similarity Index ◽  
Structural Similarity ◽  
Attention Mechanism ◽  
Generative Adversarial Networks ◽  
Reconstruction Method ◽  
Mri Imaging ◽  
Adversarial Networks ◽  
Reconstruction Methods ◽  
Mri Reconstruction

Research on undersampled magnetic resonance image (MRI) reconstruction can increase the speed of MRI imaging and reduce patient suffering. In this paper, an undersampled MRI reconstruction method based on Generative Adversarial Networks with the Self-Attention mechanism and the Relative Average discriminator (SARA-GAN) is proposed. In our SARA-GAN, the relative average discriminator theory is applied to make full use of the prior knowledge, in which half of the input data of the discriminator is true and half is fake. At the same time, a self-attention mechanism is incorporated into the high-layer of the generator to build long-range dependence of the image, which can overcome the problem of limited convolution kernel size. Besides, spectral normalization is employed to stabilize the training process. Compared with three widely used GAN-based MRI reconstruction methods, i.e., DAGAN, DAWGAN, and DAWGAN-GP, the proposed method can obtain a higher peak signal-to-noise ratio (PSNR) and structural similarity index measure(SSIM), and the details of the reconstructed image are more abundant and more realistic for further clinical scrutinization and diagnostic tasks.

QUALITY MEASURE OF THE COMPRESSED ECHO, X-RAY AND CT IMAGES

2013 ◽  
Vol 13 (01) ◽  
pp. 1350006 ◽  
Author(s):  
RAJANI GUPTA ◽  
PRASHANT BANSOD ◽  
R. S. GAMAD
Keyword(s):  
Signal To Noise Ratio ◽  
Similarity Index ◽  
Quality Measurement ◽  
Structural Similarity ◽  
Set Partitioning ◽  
Underlying Structure ◽  
Biomedical Images ◽  
X Ray ◽  
Index Measurement

The paper reveals the analysis of the compression quality of true color medical images of echocardiogram (ECHO), X-radiation (X-ray) and computed tomography (CT) and further a comparison of compressed biomedical images of various sizes using two lossy compression techniques, set partitioning in hierarchical trees (SPIHT) and discrete cosine transform (DCT) to the original image is carried out. The study also evaluates the results after analyzing various objective parameters associated with the image. The objective of this analysis is to exhibits the effect of compression ratio on absolute average difference (AAD), cross correlation (CC), image fidelity (IF), mean square error (MSE), peak signal to noise ratio (PSNR) and structural similarity index measurement (SSIM) of the compressed image by SPIHT and DCT compression techniques. The results signify that the quality of the compressed image depends on resolution of the underlying structure where CT is found to be better than other image modalities. The X-ray compression results are equivalent by both the techniques. The compression results for large size biomedical images by SPIHT signifies that ECHO having comparable results to CT and X-ray while their DCT results are substandard. The compression results for comparatively smaller images of ECHO are not as good as X-ray and CT by both the compression techniques. The quality measurement of the compressed image has been designed using MATLAB.

scholarly journals Deep Learning-Based Fast TOF-PET Image Reconstruction Using Direction Information

2021 ◽  
Author(s):  
Kibo Ote ◽  
Fumio Hashimoto
Keyword(s):  
Deep Learning ◽  
Image Reconstruction ◽  
Image Quality ◽  
Signal To Noise Ratio ◽  
Structural Similarity ◽  
Fast Time ◽  
Reconstruction Method ◽  
Direction Information ◽  
Pet Image Reconstruction ◽  
Tof Pet

Abstract Deep learning has attracted attention for positron emission tomography (PET) image reconstruction task, however, it remains necessary to further improve the image quality. In this study, we propose a novel CNN-based fast time-of-flight PET (TOF-PET) image reconstruction method to fully utilize the direction information of coincidence events. The proposed method inputs view-grouped histo-images into a 3D CNN as a multi-channel image to use the direction information of coincidence events. We evaluated the proposed method using Monte Carlo simulation data obtained from a digital brain phantom. Compared to the case without it, when using direction information, the peak signal-to-noise ratio and structural similarity were improved by 1.2 dB and 0.02, at a coincidence time resolution of 300 ps. The calculation times of the proposed method were significantly faster than the conventional iterative reconstruction. These results indicate that the proposed method improves both the speed and image quality of TOF-PET image reconstruction.

Split-Bregman Algorithm with Attenuation Correction for L-Shell Polychromatic X-ray Fluorescence Computed Tomography

2020 ◽  
Vol 10 (3) ◽  
pp. 572-578
Author(s):  
Long Liu ◽  
Bo Ma ◽  
Xiaowei Liu ◽  
Jinlan Liu ◽  
Ning Ding
Keyword(s):  
Computed Tomography ◽  
Attenuation Correction ◽  
High Efficiency ◽  
Projection Data ◽  
Reconstruction Method ◽  
X Rays ◽  
Split Bregman ◽  
X Ray ◽  
Early Cancer Diagnosis ◽  
Split Bregman Algorithm

L-shell X-ray fluorescence computed tomography based on polychromatic X-rays is a promising imaging technique for early cancer diagnosis. However, the presence of self-absorption and the long scanning time limit its usage in clinic. In this work, a reconstruction method based on split-Bregman algorithm which used sparseview projection data was proposed. Furthermore, the attenuation effect was also considered in the algorithm. In the attenuation correction, factors including the X-ray energy and the platinum concentration were taken into account. Then weighted factors calculated in the procedure of attenuation correction were added into the contribution function of pixels in the split-Bregman based reconstruction method. In the end, the feasibility of this method was tested using a cylindrical phantom with 8 mm in diameter by the Monte Carlo simulation. The phantom contained four inserts, all of which were 1.5 millimeter in diameter and filled with 0.10%, 0.20%, 0.40% and 0.80% platinum solutions, respectively. The results show that both the contrast-to-noise ratios and lowest detectable sensitivities are improved for the proposed method, comparing to the conventional MLEM. The contrast-to-noise ratios of images reconstructed by our method with 45 projections are already better than that reconstructed by MLEM with 60 projections. When using 60 projections in our method and comparing to 60 projections in the MLEM with correction, the contrast-to-noise ratio of the insert filled with 0.10% platinum solutions increased from 6.49 to 36.90, indicating its high efficiency and robustness.

scholarly journals Low-Noise Dynamic Reconstruction for X-Ray Tomographic Perfusion Studies Using Low Sampling Rates

2009 ◽  
Vol 2009 ◽  
pp. 1-13 ◽  
Author(s):  
Pau Montes ◽  
Günter Lauritsch
Keyword(s):  
Temporal Resolution ◽  
Signal To Noise Ratio ◽  
Small Animal ◽  
Low Noise ◽  
Projection Data ◽  
Interpolation Scheme ◽  
X Ray ◽  
Perfusion Studies ◽  
Temporal Smoothing ◽  
Temporal Interpolation

Functional imaging based on tomographic X-ray imaging relies on the reconstruction of a temporal sequence of images which accurately reproduces the time attenuation curves of the tissue. The main constraints of these techniques are temporal resolution and dose. Using current techniques the data acquisition has to be performed fast so that the dynamic attenuation values can be regarded as static during the scan. Due to the relatively high number of repeated scans the dose per single scan has to be low yielding a poor signal-to-noise ratio (SNR) in the reconstructed images. In a previous publication a temporal interpolation scheme in the projection data space was relaxing the temporal resolution constraint. The aim of this contribution is the improvement of the SNR. A temporal smoothing term is introduced in the temporal interpolation scheme such that only the physiologic relevant bandwidth is considered. A significant increase of the SNR is achieved. The obtained level of noise only depends on the total dose applied and is independent of the number of scans and the SNR of a single reconstructed image. The approach might be the first step towards using slowly rotating CT systems for perfusion imaging like C-arm or small animal CT scanners.

DEVELOPMENT OF AN IMAGE RECONSTRUCTION METHOD FOR MICRON-CT USING PIXE

2005 ◽  
Vol 15 (03n04) ◽  
pp. 195-202 ◽  
Author(s):  
T. YAMAGUCHI ◽  
K. ISHII ◽  
H. YAMAZAKI ◽  
S. MATSUYAMA ◽  
Y. WATANABE ◽  
...  
Keyword(s):  
Image Reconstruction ◽  
Em Algorithm ◽  
Computation Time ◽  
Spot Size ◽  
Biological Research ◽  
Projection Data ◽  
Reconstruction Method ◽  
Ct Image ◽  
X Ray ◽  
Image Reconstruction Method

A prototype of micron-CT for biological research is being developed at Tohoku University. This micron-CT uses a point X-ray source with a spot size of 1μm and an X-ray CCD with 1000×1000 pixels of 8μm×8μm, achieving a spatial resolutions of the order of micro-meter. The event data obtained by the X-ray CCD is statistically poor and the 3 dimensional filtered back projection (3D FBP) algorithm, generally used in image reconstruction of X-ray CT, is not suitable because it is highly sensitive to statistical noise. Hence, we applied the expectation maximization (EM) algorithm for image reconstruction and developed an image reconstruction method using 3D EM algorithm. To confirm the validity of the reconstruction method, we irradiated two hairs inside a micro tube and reconstructed the CT image applying both EM and FBP algorithm on projection data. With 200×200×200 voxels of 4μm×4μm×4μm in the field of view, the computation time was less than 2 mins per iteration on a DELL Precision 650 Workstation 3.2GHz. The resulting EM image showed a better contrast than FBP image, and in the EM reconstructed CT image, we were able to reconstruct the micro tube of 270μm diameter and 45μm wall thickness and to visualize the two hairs inside.

An Electron Diffraction Study on Litnerized Potato Starch

1990 ◽  
Vol 48 (1) ◽  
pp. 580-581
Author(s):  
Jean-Claude Jésior ◽  
Roger Vuong ◽  
Henri Chanzy
Keyword(s):  
Electron Microscope ◽  
Electron Diffraction ◽  
Signal To Noise Ratio ◽  
Potato Starch ◽  
Image Intensifier ◽  
Electron Diffraction Study ◽  
Starch Granules ◽  
X Ray ◽  
Individual Grains ◽  
Diamond Knife

Starch is arranged in a crystalline manner within its storage granules and should thus give sharp X-ray diagrams. Unfortunately most of the common starch granules have sizes between 1 and 100μm, making them too small for an X-ray study on individual grains. There is only one instance where an oriented X-ray diagram could be obtained on one sector of an individual giant starch granule. Despite their small size, starch granules are still too thick to be studied by electron diffraction with a transmission electron microscope. The only reported study on starch ultrastructure using electron diffraction on frozen hydrated material was made on small fragments. The present study has been realized on thin sectioned granules previously litnerized to improve the signal to noise ratio.Potato starch was hydrolyzed for 10 days in 2.2N HCl at 35°C, dialyzed against water until neutrality and embedded in Nanoplast. Sectioning was achieved with a commercially available low-angle “35°” diamond knife (Diatome) after a very carefull trimming and a pre-sectioning with a classical “45°” diamond knife. Sections obtained at a final sectioning angle of 42.2° (compared with the usual 55-60°) and at a nominal thickness of 900Å were collected on a Formvar-carbon coated grid. The exact location of the starch granules in their sections was recorded by optical microscopy on a Zeiss Universal polarizing microscope (Fig. 1a). After rehydration at a relative humidity of 95% for 24 hours they were mounted on a Philips cryoholder and quench frozen in liquid nitrogen before being inserted under frozen conditions in a Philips EM 400T electron microscope equipped with a Gatan anticontaminator and a Lhesa image intensifier.

Nondestructive Inspection of Thin, Low-Z Samples using Multiplexed Compton Scatter Tomography

1997 ◽  
Vol 503 ◽  
Author(s):  
B. L. Evans ◽  
J. B. Martin ◽  
L. W. Burggraf
Keyword(s):  
Signal To Noise Ratio ◽  
Reconstruction Algorithm ◽  
Filtered Backprojection ◽  
Line Broadening ◽  
Compton Scatter ◽  
X Ray ◽  
Tomography System ◽  
Transmission Computed Tomography ◽  
High Purity Germanium ◽  
Scattered Energy

ABSTRACTThe viability of a Compton scattering tomography system for nondestructively inspecting thin, low Z samples for corrosion is examined. This technique differs from conventional x-ray backscatter NDI because it does not rely on narrow collimation of source and detectors to examine small volumes in the sample. Instead, photons of a single energy are backscattered from the sample and their scattered energy spectra are measured at multiple detector locations, and these spectra are then used to reconstruct an image of the object. This multiplexed Compton scatter tomography technique interrogates multiple volume elements simultaneously. Thin samples less than 1 cm thick and made of low Z materials are best imaged with gamma rays at or below 100 keV energy. At this energy, Compton line broadening becomes an important resolution limitation. An analytical model has been developed to simulate the signals collected in a demonstration system consisting of an array of planar high-purity germanium detectors. A technique for deconvolving the effects of Compton broadening and detector energy resolution from signals with additive noise is also presented. A filtered backprojection image reconstruction algorithm with similarities to that used in conventional transmission computed tomography is developed. A simulation of a 360–degree inspection gives distortion-free results. In a simulation of a single-sided inspection, a 5 mm × 5 mm corrosion flaw with 50% density is readily identified in 1-cm thick aluminum phantom when the signal to noise ratio in the data exceeds 28.

A Spectrum-Adaptive Decomposition Method for Effective Atomic Number Estimation Using Dual Energy CT

2020 ◽  
Vol 2020 (14) ◽  
pp. 293-1-293-7
Author(s):  
Ankit Manerikar ◽  
Fangda Li ◽  
Avinash C. Kak
Keyword(s):  
Atomic Number ◽  
Decomposition Method ◽  
Traditional Approach ◽  
Effective Atomic Number ◽  
Dual Energy ◽  
Performance Comparison ◽  
Estimation Methods ◽  
Dual Energy Ct ◽  
Projection Data ◽  
X Ray

Dual Energy Computed Tomography (DECT) is expected to become a significant tool for voxel-based detection of hazardous materials in airport baggage screening. The traditional approach to DECT imaging involves collecting the projection data using two different X-ray spectra and then decomposing the data thus collected into line integrals of two independent characterizations of the material properties. Typically, one of these characterizations involves the effective atomic number (Zeff) of the materials. However, with the X-ray spectral energies typically used for DECT imaging, the current best-practice approaches for dualenergy decomposition yield Zeff values whose accuracy range is limited to only a subset of the periodic-table elements, more specifically to (Z < 30). Although this estimation can be improved by using a system-independent ρe — Ze (SIRZ) space, the SIRZ transformation does not efficiently model the polychromatic nature of the X-ray spectra typically used in physical CT scanners. In this paper, we present a new decomposition method, AdaSIRZ, that corrects this shortcoming by adapting the SIRZ decomposition to the entire spectrum of an X-ray source. The method reformulates the X-ray attenuation equations as direct functions of (ρe, Ze) and solves for the coefficients using bounded nonlinear least-squares optimization. Performance comparison of AdaSIRZ with other Zeff estimation methods on different sets of real DECT images shows that AdaSIRZ provides a higher output accuracy for Zeff image reconstructions for a wider range of object materials.

Sign in / Sign up

Export Citation Format

Share Document