Relative Entropy of Correct Proximal Policy Optimization Algorithms with Modified Penalty Factor in Complex Environment.

In the field of reinforcement learning, we propose a Correct Proximal Policy Optimization (CPPO) algorithm based on the modified penalty factor ß and relative entropy in order to solve the robustness and stationarity of traditional algorithms. Firstly, In the process of reinforcement learning, this paper establishes a strategy evaluation mechanism through the policy distribution function. Secondly, the state space function is quantified by introducing entropy, whereby the approximation policy is used to approximate the real policy distribution, and the kernel function estimation and calculation of relative entropy is used to fit the reward function based on complex problem. Finally, through the comparative analysis on the classic test cases, we demonstrated that our proposed algorithm is effective, has a faster convergence speed and better performance than the traditional PPO algorithm, and the measure of the relative entropy can show the differences. In addition, it can more efficiently use the information of complex environment to learn policies. At the same time, not only can our paper explain the rationality of the policy distribution theory, the proposed framework can also balance between iteration steps, computational complexity and convergence speed, and we also introduced an effective measure of performance using the relative entropy concept.

Explainable coronary artery disease prediction model based on AutoGluon from AutoML framework.

Objective: This study focuses on the innovative application of Automated Machine Learning (AutoML) technology in cardiovascular medicine to construct an explainable Coronary Artery Disease (CAD) prediction model to support the clinical diagnosis of CAD. Methods: This study utilizes a combined data set of five public data sets related to CAD. An ensemble model is constructed using the AutoML open-source framework AutoGluon to evaluate the feasibility of AutoML in constructing a disease prediction model in cardiovascular medicine. The performance of the ensemble model is compared against individual baseline models. Finally, the disease prediction ensemble model is explained using SHapley Additive exPlanations (SHAP). Results: The experimental results show that the AutoGluon-based ensemble model performs better than the individual baseline models in predicting CAD. It achieved an accuracy of 0.9167 and an AUC of 0.9562 in 4-fold cross-bagging. SHAP measures the importance of each feature to the prediction of the model and explains the prediction results of the model. Conclusion: This study demonstrates the feasibility and efficacy of AutoML technology in cardiovascular medicine and highlights its potential in disease prediction. AutoML reduces the barriers to model building and significantly improves prediction accuracy. Additionally, the integration of SHAP enhances model transparency and explainability, which is critical to ensuring model credibility and widespread adoption in cardiovascular medicine.

Identification of Ultrasonic Echolucent Carotid Plaques Using Discrete Fréchet Distance Between Bimodal Gamma Distributions.

OBJECTIVE: Echolucent carotid plaques are associated with acute cardiovascular and cerebrovascular events (ACCEs) in atherosclerotic patients. The aim of this study was to develop a computer-aided method for identifying echolucent plaques. METHODS: A total of 315 ultrasound images of carotid plaques (105 echo-rich, 105 intermediate, and 105 echolucent) collected from 153 patients were included in this study. A bimodal gamma distribution was proposed to model the pixel statistics in the gray scale images of plaques. The discrete Fréchet distance features (DFDFs) of each plaque were extracted based on the statistical model. The most discriminative features (MDFs) were obtained from DFDFs by the linear discriminant analysis, and a k-nearest-neighbor classifier was implemented for classification of different types of plaques. RESULTS: The classification accuracy of the three types of plaques using MDFs can reach 77.46%. When a receiver operating characteristics curve was produced to identify echolucent plaques, the area under the curve was 0.831. CONCLUSION: Our results indicate potential feasibility of the method for identifying echolucent plaques based on DFDFs. SIGNIFICANCE: Our method may potentially improve the ability of noninvasive ultrasonic examination in risk prediction of ACCEs for patients with plaques.

Peroneal perforator pedicle propeller flap for lower leg soft tissue defect reconstruction: Clinical applications and treatment of venous congestion.

Objective To describe the characteristics of the perforator vessel in the peroneal artery of the lower leg and to explore the use of perforator pedicle propeller flaps to repair soft tissue defects in the lower leg, heel and foot. Methods This retrospective study enrolled patients with soft tissue defects of the distal lower leg, heel and foot who underwent surgery using peroneal perforator-based propeller flaps. The peroneal artery perforators were identified preoperatively by colour duplex Doppler ultrasound. The flap was designed based on the preoperatively-identified perforator location, with the posterior border of the fibula employed as an axis, and the perforator vessel as the pivot point of rotation. Patients were followed-up to determine the outcomes. Results The study analysed 36 patients (mean age, 39.7 years). The majority of the soft tissue defects were on the heel (20; 55.6%). The donor-site of the flap was closed in 11 patients by direct suturing and skin grafting was undertaken in 25 patients. Postoperative complications included venous congestion (nine patients), which was managed with delayed wound coverage and bleeding therapy. All wounds were eventually cured and the flaps were cosmetically acceptable. Conclusions The peroneal perforator pedicle propeller flap is an appropriate choice to repair soft tissue defects of the distal limbs.

Reconstruction for 3D PET Based on Total Variation Constrained Direct Fourier Method.

This paper presents a total variation (TV) regularized reconstruction algorithm for 3D positron emission tomography (PET). The proposed method first employs the Fourier rebinning algorithm (FORE), rebinning the 3D data into a stack of ordinary 2D data sets as sinogram data. Then, the resulted 2D sinogram are ready to be reconstructed by conventional 2D reconstruction algorithms. Given the locally piece-wise constant nature of PET images, we introduce the total variation (TV) based reconstruction schemes. More specifically, we formulate the 2D PET reconstruction problem as an optimization problem, whose objective function consists of TV norm of the reconstructed image and the data fidelity term measuring the consistency between the reconstructed image and sinogram. To solve the resulting minimization problem, we apply an efficient methods called the Bregman operator splitting algorithm with variable step size (BOSVS). Experiments based on Monte Carlo simulated data and real data are conducted as validations. The experiment results show that the proposed method produces higher accuracy than conventional direct Fourier (DF) (bias in BOSVS is 70% of ones in DF, variance of BOSVS is 80% of ones in DF).

Patient-specific geometrical modeling of orthopedic structures with high efficiency and accuracy for finite element modeling and 3D printing.

We improved the geometrical modeling procedure for fast and accurate reconstruction of orthopedic structures. This procedure consists of medical image segmentation, three-dimensional geometrical reconstruction, and assignment of material properties. The patient-specific orthopedic structures reconstructed by this improved procedure can be used in the virtual surgical planning, 3D printing of real orthopedic structures and finite element analysis. A conventional modeling consists of: image segmentation, geometrical reconstruction, mesh generation, and assignment of material properties. The present study modified the conventional method to enhance software operating procedures. Patient's CT images of different bones were acquired and subsequently reconstructed to give models. The reconstruction procedures were three-dimensional image segmentation, modification of the edge length and quantity of meshes, and the assignment of material properties according to the intensity of gravy value. We compared the performance of our procedures to the conventional procedures modeling in terms of software operating time, success rate and mesh quality. Our proposed framework has the following improvements in the geometrical modeling: (1) processing time: (femur: 87.16 ± 5.90 %; pelvis: 80.16 ± 7.67 %; thoracic vertebra: 17.81 ± 4.36 %; P < 0.05); (2) least volume reduction (femur: 0.26 ± 0.06 %; pelvis: 0.70 ± 0.47, thoracic vertebra: 3.70 ± 1.75 %; P < 0.01) and (3) mesh quality in terms of aspect ratio (femur: 8.00 ± 7.38 %; pelvis: 17.70 ± 9.82 %; thoracic vertebra: 13.93 ± 9.79 %; P < 0.05) and maximum angle (femur: 4.90 ± 5.28 %; pelvis: 17.20 ± 19.29 %; thoracic vertebra: 3.86 ± 3.82 %; P < 0.05). Our proposed patient-specific geometrical modeling requires less operating time and workload, but the orthopedic structures were generated at a higher rate of success as compared with the conventional method. It is expected to benefit the surgical planning of orthopedic structures with less operating time and high accuracy of modeling.

Pulsatile flow characterization in a vessel phantom with elastic wall using ultrasonic particle image velocimetry technique: the impact of vessel stiffness on flow dynamics.

This study aims to experimentally investigate the impact of vessel stiffness on the flow dynamics of pulsatile vascular flow. Vessel phantoms with elastic walls were fabricated using polyvinyl alcohol cryogel to result in stiffness ranging from 60.9 to 310.3 kPa and tested with pulsatile flows using a flow circulation set-up. Two-dimensional instantaneous and time-dependent flow velocity and shear rate vector fields were measured using ultrasonic particle image velocimetry (EchoPIV). The waveforms of peak velocities measured by EchoPIV were compared with the ultrasonic pulse Doppler spectrum, and the measuring accuracy was validated. The cyclic vessel wall motion and flow pressure were obtained as well. The results showed that vessel stiffening influenced the waveforms resulting from vessel wall distension and flow pressure, and the fields of flow velocity and shear rate. The stiffer vessel had smaller inner diameter variation, larger pulse pressure and median pressure. The velocity and shear rate maximized at peak systole for all vessels. The results showed a decrease in wall shear stress for a stiffer vessel, which can initiate the atherosclerotic process. Our study elucidates the impact of vessel stiffness on several flow dynamic parameters, and also demonstrates the EchoPIV technique to be a useful and powerful tool in cardiovascular research.

Investigation of atrial vulnerability by analysis of the sinus node EG from atrial fibrillation models using a phase synchronization method.

Atrial fibrillation (AF) can result in life-threatening arrhythmia, and a clinically convenient means for detecting vulnerability remains elusive. We investigated atrial vulnerability by analyzing the sinus electrogram (EG) from AF animal models using a phase synchronization method. Using acetylcholine (ACh)-induced acute canine AF models (n= 4), a total of 128 electrical leads were attached to the surface of the anterior and posterior atria, and the pulmonary veins to form an electrocardiological mapping system. ACh was injected at varying concentrations with ladder-type adjustments. Sinus EGs and induced AF EGs that pertain to specific ACh concentrations were recorded.We hypothesize that the atrial vulnerability may be correlated with the Shannon entropy (SE) of the phase difference matrix that is extracted from the sinus EG. Our research suggests that the combination of SE with the synchronization method enables the sinus node EG to be analyzed and used to estimate atrial vulnerability.

Automatic target recognition based on cross-plot.

Automatic target recognition that relies on rapid feature extraction of real-time target from photo-realistic imaging will enable efficient identification of target patterns. To achieve this objective, Cross-plots of binary patterns are explored as potential signatures for the observed target by high-speed capture of the crucial spatial features using minimal computational resources. Target recognition was implemented based on the proposed pattern recognition concept and tested rigorously for its precision and recall performance. We conclude that Cross-plotting is able to produce a digital fingerprint of a target that correlates efficiently and effectively to signatures of patterns having its identity in a target repository.

A geometrical perspective for the bargaining problem.

A new treatment to determine the Pareto-optimal outcome for a non-zero-sum game is presented. An equilibrium point for any game is defined here as a set of strategy choices for the players, such that no change in the choice of any single player will increase the overall payoff of all the players. Determining equilibrium for multi-player games is a complex problem. An intuitive conceptual tool for reducing the complexity, via the idea of spatially representing strategy options in the bargaining problem is proposed. Based on this geometry, an equilibrium condition is established such that the product of their gains over what each receives is maximal. The geometrical analysis of a cooperative bargaining game provides an example for solving multi-player and non-zero-sum games efficiently.

Numerical simulation and experimental validation of swirling flow in spiral vortex ventricular assist device.

Spiral vortex ventricular assist device (SV-VAD) supports cardiac patients with refractory heart failure. Unfortunately, thrombus formation and risk of stroke due to flow complications may lead to aggravated conditions. The hemodynamics of a continuous flow in the ventricular chamber of a SV-VAD can be analyzed using numerical simulation. Particle image velocimetry and laser Doppler anemometry are utilized for validating the simulated spiral flow in a transparent acrylic SV-VAD replica based on its cross-sectional averaged axial and tangential velocities. After validation, the relationship between swirling flow and blood cell damage is established by evaluating flow effect on thrombosis due to high shear stress. Based on our analysis, stagnancy of the flow within the SV-VAD is insignificant and its low shear stress minimizes hemolysis.

Noninvasive cardiac flow assessment using high speed magnetic resonance fluid motion tracking.

Cardiovascular diseases can be diagnosed by assessing abnormal flow behavior in the heart. We introduce, for the first time, a magnetic resonance imaging-based diagnostic that produces sectional flow maps of cardiac chambers, and presents cardiac analysis based on the flow information. Using steady-state free precession magnetic resonance images of blood, we demonstrate intensity contrast between asynchronous and synchronous proton spins. Turbulent blood flow in cardiac chambers contains asynchronous blood proton spins whose concentration affects the signal intensities that are registered onto the magnetic resonance images. Application of intensity flow tracking based on their non-uniform signal concentrations provides a flow field map of the blood motion. We verify this theory in a patient with an atrial septal defect whose chamber blood flow vortices vary in speed of rotation before and after septal occlusion. Based on the measurement of cardiac flow vorticity in our implementation, we establish a relationship between atrial vorticity and septal defect. The developed system has the potential to be used as a prognostic and investigative tool for assessment of cardiac abnormalities, and can be exploited in parallel to examining myocardial defects using steady-state free precession magnetic resonance images of the heart.

Theory and validation of magnetic resonance fluid motion estimation using intensity flow data.

BACKGROUND: Motion tracking based on spatial-temporal radio-frequency signals from the pixel representation of magnetic resonance (MR) imaging of a non-stationary fluid is able to provide two dimensional vector field maps. This supports the underlying fundamentals of magnetic resonance fluid motion estimation and generates a new methodology for flow measurement that is based on registration of nuclear signals from moving hydrogen nuclei in fluid. However, there is a need to validate the computational aspect of the approach by using velocity flow field data that we will assume as the true reference information or ground truth. METHODOLOGY/PRINCIPAL FINDINGS: In this study, we create flow vectors based on an ideal analytical vortex, and generate artificial signal-motion image data to verify our computational approach. The analytical and computed flow fields are compared to provide an error estimate of our methodology. The comparison shows that the fluid motion estimation approach using simulated MR data is accurate and robust enough for flow field mapping. To verify our methodology, we have tested the computational configuration on magnetic resonance images of cardiac blood and proved that the theory of magnetic resonance fluid motion estimation can be applicable practically. CONCLUSIONS/SIGNIFICANCE: The results of this work will allow us to progress further in the investigation of fluid motion prediction based on imaging modalities that do not require velocity encoding. This article describes a novel theory of motion estimation based on magnetic resonating blood, which may be directly applied to cardiac flow imaging.