Three robust edges stopping functions for image denoising

In this paper, we present three strong edge stopping functions for image enhancement. These edge stopping functions have the advantage of effectively removing the image noise while preserving the true edges and other important features. The obtained results show an improved quality for the restored images compared to existing restoration models.

A Study on Distortion Estimation Based on Image Gradients

Image noise is a variation of uneven pixel values that occurs randomly. A good estimation of image noise parameters is crucial in image noise modeling, image denoising, and image quality assessment. To the best of our knowledge, there is no single estimator that can predict all noise parameters for multiple noise types. The first contribution of our research was to design a noise data feature extractor that can effectively extract noise information from the image pair. The second contribution of our work leveraged other noise parameter estimation algorithms that can only predict one type of noise. Our proposed method, DE-G, can estimate additive noise, multiplicative noise, and impulsive noise from single-source images accurately. We also show the capability of the proposed method in estimating multiple corruptions.

Visualization of bone details in a novel photon-counting dual-source CT scanner—comparison with energy-integrating CT

Objectives Photon-counting detector CT (PCD-CT) promises a leap in spatial resolution due to smaller detector pixel sizes than implemented in energy-integrating detector CTs (EID-CT). Our objective was to compare the visualization of smallest bone details between PCD-CT and EID-CT using a mouse as a specimen. Materials and methods Two euthanized mice were scanned at a 20-slice EID-CT and a dual-source PCD-CT in single-pixel mode at various CTDIVol values. Image noise and signal-to-noise ratio (SNR) were evaluated using repeated ROI measurements. Edge sharpness of bones was compared by the maximal slope within CT value plots along sampling lines intersecting predefined bones of the spine. Two readers evaluated bone detail visualization at four regions of the spine on a three-point Likert scale at various CTDIVol’s. Two radiologists selected the series with better detail visualization among each of 20 SNR-matched pairs of EID-CT and PCD-CT series. Results In CTDIVol-matched scans, PCD-CT series showed significantly lower image noise (NoiseCTDI=5 mGy: 16.27 ± 1.39 vs. 23.46 ± 0.96 HU, p < 0.01), higher SNR (SNRCTDI=5 mGy: 20.57 ± 1.89 vs. 14.00 ± 0.66, p < 0.01), and higher edge sharpness (Edge Slopelumbar spine: 981 ± 160 vs. 608 ± 146 HU/mm, p < 0.01) than EID-CT series. Two radiologists considered the delineation of bone details as feasible at consistently lower CTDIVol values at PCD-CT than at EID-CT. In comparison of SNR-matched reconstructions, PCD-CT series were still considered superior in almost all cases. Conclusions In this head-to-head comparison, PCD-CT showed superior objective and subjective image quality characteristics over EID-CT for the delineation of tiniest bone details. Even in SNR-matched pairs (acquired at different CTDIVol’s), PCD-CT was strongly preferred by radiologists. Key Points • In dose-matched scans, photon-counting detector CT series showed significantly less image noise, higher signal-to-noise ratio, and higher edge sharpness than energy-integrating detector CT series. • Human observers considered the delineation of tiny bone details as feasible at much lower dose levels in photon-counting detector CT than in energy-integrating detector CT. • In direct comparison of series matched for signal-to-noise ratio, photon-counting detector CT series were considered superior in almost all cases.

Automatic Recognition and Correction of Volleyball Players’ Release Angle Based on Feature Statistics

The constant reform of the competition rules has promoted the innovation of volleyball techniques and tactics. In order to improve the training efficiency and competitive level of volleyball players, this study designed a volleyball player shooting angle automatic recognition and correction method based on the process of feature statistics. Firstly, the basic structure of the information acquisition system is analyzed, and the acquisition process is determined. Then, grayscale and binarization operations are carried out for color-moving images to separate their foreground and background, and a median filtering algorithm is used to remove the image noise. Then, the image pyramid of different sizes is generated by the filter. Based on setting the datum direction, the feature of volleyball shooting is extracted by using the line formula. On this basis, we construct a support vector machine (SVM) classifier to statistically classify the features, use the histogram additive kernel support vector machine method to obtain the lens angle recognition results, and correct the lens angle through feature point matching. Simulation experiments show that this method can effectively remove image noise and make the image signal-to-noise ratio higher, and it can effectively identify whether volleyball players’ release Angle is correct, to achieve the purpose of timely correction.

Improving robustness of a deep learning-based lung-nodule classification model of CT images with respect to image noise

Objective: Robustness is an important aspect to consider, when developing methods for medical image analysis. This study investigated robustness properties of deep neural networks (DNNs) for a lung nodule classification problem based on CT images and proposed a solution to improve robustness. Approach: We firstly constructed a class of four DNNs with different widths, each predicting an output label (benign or malignant) for an input CT image cube containing a lung nodule. These networks were trained to achieve Area Under the Curve of 0.891-0.914 on a testing dataset. We then added to the input CT image cubes noise signals generated randomly using a realistic CT image noise model based on a noise power spectrum at 100 mAs, and monitored the DNN’s output change. We defined $SAR_{5} (\%)$ to quantify the robustness of the trained DNN model, indicating that for $5\%$ of CT image cubes, the noise can change the prediction results with a chance of at least $SAR_{5} (\%)$. To understand robustness, we viewed the information processing pipeline by the DNN as a two-step process, with the first step using all but the last layers to extract representations of the input CT image cubes in a latent space, and the second step employing the last fully-connected layer as a linear classifier to determine the position of the sample representations relative to a decision plane. To improve robustness, we proposed to retrain the last layer of the DNN with a Supporting Vector Machine (SVM) hinge loss function to enforce the desired position of the decision plane. Main results: $SAR_{5}$ ranged in $47.0\sim 62.0\%$ in different DNNs. The unrobustness behavior may be ascribed to the unfavorable placement of the decision plane in the latent representation space, which made some samples be perturbed to across the decision plane and hence susceptible to noise. The DNN-SVM model improved robustness over the DNN model and reduced $SAR_{5}$ by $8.8\sim 21.0\%$. Significance: This study provided insights about the potential reason for the unrobustness behavior of DNNs and the proposed DNN-SVM model improved model robustness.

Application of deep learning image reconstruction algorithm to improve image quality in CT angiography of children with Takayasu arteritis

BACKGROUND: The inflammatory indexes of children with Takayasu arteritis (TAK) usually tend to be normal immediately after treatment, therefore, CT angiography (CTA) has become an important method to evaluate the status of TAK and sometime is even more sensitive than laboratory test results. OBJECTIVE: To evaluate image quality improvement in CTA of children diagnosed with TAK using a deep learning image reconstruction (DLIR) in comparison to other image reconstruction algorithms. METHODS: hirty-two TAK patients (9.14±4.51 years old) underwent neck, chest and abdominal CTA using 100 kVp were enrolled. Images were reconstructed at 0.625 mm slice thickness using Filtered Back-Projection (FBP), 50%adaptive statistical iterative reconstruction-V (ASIR-V), 100%ASIR-V and DLIR with high setting (DLIR-H). CT number and standard deviation (SD) of the descending aorta and back muscle were measured and contrast-to-noise ratio (CNR) for aorta was calculated. The vessel visualization, overall image noise and diagnostic confidence were evaluated using a 5-point scale (5, excellent; 3, acceptable) by 2 observers. RESULTS: There was no significant difference in CT number across images reconstructed using different algorithms. Image noise values (in HU) were 31.36±6.01, 24.96±4.69, 18.46±3.91 and 15.58±3.65, and CNR values for aorta were 11.93±2.12, 15.66±2.37, 22.54±3.34 and 24.02±4.55 using FBP, 50%ASIR-V, 100%ASIR-V and DLIR-H, respectively. The 100%ASIR-V and DLIR-H images had similar noise and CNR (all P > 0.05), and both had lower noise and higher CNR than FBP and 50%ASIR-V images (all P < 0.05). The subjective evaluation suggested that all images were diagnostic for large arteries, however, only 50%ASIR-V and DLIR-H met the diagnostic requirement for small arteries (3.03±0.18 and 3.53±0.51). CONCLUSION: DLIR-H improves CTA image quality and diagnostic confidence for TAK patients compared with 50%ASIR-V, and best balances image noise and spatial resolution compared with 100%ASIR-V.

Radiation dose reduction with deep-learning image reconstruction for coronary computed tomography angiography

Objectives Deep-learning image reconstruction (DLIR) offers unique opportunities for reducing image noise without degrading image quality or diagnostic accuracy in coronary CT angiography (CCTA). The present study aimed at exploiting the capabilities of DLIR to reduce radiation dose and assess its impact on stenosis severity, plaque composition analysis, and plaque volume quantification. Methods This prospective study includes 50 patients who underwent two sequential CCTA scans at normal-dose (ND) and lower-dose (LD). ND scans were reconstructed with Adaptive Statistical Iterative Reconstruction-Veo (ASiR-V) 100%, and LD scans with DLIR. Image noise (in Hounsfield units, HU) and quantitative plaque volumes (in mm3) were assessed quantitatively. Stenosis severity was visually categorized into no stenosis (0%), stenosis (< 20%, 20–50%, 51–70%, 71–90%, 91–99%), and occlusion (100%). Plaque composition was classified as calcified, non-calcified, or mixed. Results Reduction of radiation dose from ND scans with ASiR-V 100% to LD scans with DLIR at the highest level (DLIR-H; 1.4 mSv vs. 0.8 mSv, p < 0.001) had no impact on image noise (28 vs. 27 HU, p = 0.598). Reliability of stenosis severity and plaque composition was excellent between ND scans with ASiR-V 100% and LD scans with DLIR-H (intraclass correlation coefficients of 0.995 and 0.974, respectively). Comparison of plaque volumes using Bland–Altman analysis revealed a mean difference of − 0.8 mm3 (± 2.5 mm3) and limits of agreement between − 5.8 and + 4.1 mm3. Conclusion DLIR enables a reduction in radiation dose from CCTA by 43% without significant impact on image noise, stenosis severity, plaque composition, and quantitative plaque volume. Key Points •Deep-learning image reconstruction (DLIR) enables radiation dose reduction by over 40% for coronary computed tomography angiography (CCTA). •Image noise remains unchanged between a normal-dose CCTA reconstructed by ASiR-V and a lower-dose CCTA reconstructed by DLIR. •There is no impact on the assessment of stenosis severity, plaque composition, and quantitative plaque volume between the two scans.

Advanced Noise-Optimized Dual-Energy Virtual Monochromatic Imaging vs. Conventional 120-kVp CT Imaging: Image Quality Assessment

Purpose: This study aimed at evaluating the image quality characteristics of advanced noise-optimized and traditional virtual monochromatic images compared with conventional 120-kVp images from second-generation Dual-Source CT. Materials and Methods: For spiral scans six syringes filled with diluted iodine contrast material (1, 2, 5, 10, 15, 20 mg I/ml) were inserted into the test phantom and scanned with a second-generation dual-source CT in both single-energy (120-kVp) and dual-energy modes. Images set contain conventional single-energy 120-kVp, and virtual monochromatic were reconstructed with energies ranging from 40 to 190-keV in 1-keV steps. An energy-domain noise reduction algorithm was applied and the mean CT number, image noise, and iodine CNR were calculated. Results: The iodine CT number of conventional 120-kVp images compared with monochromatic of 40-, 50-, 60- and 70-keV images showed increase. The improvement ratio of image noise on Advanced Virtual Monochromatic Images (AVMIs) compared with the Traditional Virtual Monochromatic Images (TVMIs) at energies of 40-, 50-, 60, 70-keV was 52.9%, 35.7%, 8.1%, 2.1%, respectively. At AVMIs from 75- to 190-keV, the image noise value was less than conventional 120-kVp images. CNR improvement ratio at 20 mg/ml of iodinated contrast material for TVMIs and AVMIs compared to 120-kVp CT images and AVMIs compared to TVMI was 18.3% and 56.3%, 32.1% respectively. Conclusion: Both TVMIs (in energies ranging from 54 to 71-keV) and AVMIs (in energies ranging from 40 to 74-keV) represent improvement in the iodine contrast-to-noise ratio than conventional 120-kVp CT images for the same radiation dose. Also, AVMIs compared to TVMIs have been obtained considerable noise reduction and CNR improvement for low-energy virtual monochromatic images. In the present study, we show that virtual monochromatic image and its Advanced version (AVMI) may boost the dual-energy CT advantages by providing higher CNR images in the same exposure value compared to routinely acquired single-energy CT images.

Studies on the radiation dose, image quality and low contrast detectability from MSCT abdomen by using low tube voltage

Background To study radiation dose, image quality and low-contrast cylinder detectability from multislice CT (MSCT) abdomen by using low tube voltage using the American College of Radiology (ACR) phantom. The ACR phantom (low-contrast module) was scanned with 64 MSCT scanner (Brilliance, Philips Medical System, Eindhoven, Netherlands) with 80 and 120 KVP, utilizing different tube current time product (mAs) range from 50 to 380 mAs. The image noise (SD), signal to noise ratio, contrast-to-noise ratio (CNR), and scores of low contrast detectability were assessed for every image respectively. Results From images analyses, the noise essentially increased with the use of low tube voltage. The CNR was 0.94 ± 0.27 at 120 KVP, and CNR was 0.43 ± 0.22 at 80 KVP. However, with the same dose, there were no differences of statistical significance in scores of low-contrast detectability between 120 KVP at 300mAs and 80 KVP at (200–380) mAs (p > 0.05). At 300 mAs, the CTDIvol obtained at 80 KVP was about 29% of that at 120 KVP. The CTDIvol obtained at 80 KVP were decreased from 5% at 50 mAs, to 37% at 380 mAs. Conclusions There is a possibility to decrease exposure of radiation virtually by reducing KVP from 120 to 80 KVP in examination of abdominal CT when the high tube current is used, though increasing image noise at low tube voltage.