Based on the imaging photoplethysmography (iPPG) and blind source separation (BSS) theory the author put forward a method for non-contact heartbeat frequency estimation. Using the recorded video images of the human face in the ambient light with Webcam, we detected the human face through software, separated the detected facial image into three channels RGB components. And then preprocesses i.e. normalization, whitening, etc. were carried out to a certain number of RGB data. After the independent component analysis (ICA) theory and joint approximate diagonalization of eigenmatrices (JADE) algorithm were applied, we estimated the frequency of heart rate through spectrum analysis. Taking advantage of the consistency of Bland-Altman theory analysis and the commercial Pulse Oximetry Sensor test results, the root mean square error of the algorithm result was calculated as 2.06 beat/min. It indicated that the algorithm could realize the non-contact measurement of heart rate and lay the foundation for the remote and non-contact measurement of multi-parameter physiological measurements.
Heart rate is a crucial indicator of human health with significant physiological importance. Traditional contact methods for measuring heart rate, such as electrocardiograph or wristbands, may not always meet the need for convenient health monitoring. Remote photoplethysmography (rPPG) provides a non-contact method for measuring heart rate and other physiological indicators by analyzing blood volume pulse signals. This approach is non-invasive, does not require direct contact, and allows for long-term healthcare monitoring. Deep learning has emerged as a powerful tool for processing complex image and video data, and has been increasingly employed to extract heart rate signals remotely. This article reviewed the latest research advancements in rPPG-based heart rate measurement using deep learning, summarized available public datasets, and explored future research directions and potential advancements in non-contact heart rate measurement.
The velocity of blood in vessels is an important indicator that reflects the microcirculatory status. The core of the measurement technology, which is based on spatiotemporal (ST) image, is to map the cell motion trace to the two-dimensional ST image, and transfer the measurement of flow velocity to the detection of trace orientation in ST image. This paper proposes a trace orientation measurement algorithm is based on Randomized Hough Transformation and projection transformation, and it is able to estimate trace orientation and flow velocity in noisy ST images. Experiments showed that the agreement between the results by manual and by the proposed algorithm reached over 90%.
Stage ⅠA lung adenocarcinoma presented as ground glass dominant on thin-section high-resolution CT scan is a special subtype of lung cancer. The characteristics of this subtype are quite different from the other patients, which presented as lower malignancy and better prognosis. Clinical, pathological and imaging studies have revealed that the proportion of the solid component in part-solid ground glass nodule is closely related with the pathological type and the prognosis of lung cancer. The methods for the assessment of the solid components in the ground glass nodule can be classified into three types, including subjective assessment, two dimensional measurement and three dimensional measurement. This review summarized the advantages and the limitations of these three methods. We also reviewed the clinical application of these techniques.
ObjectiveAntiviral treatments could benefit chronic hepatitis B (CHB) patients with the regression or improvement of liver fibrosis. However, the degree of dynamic change of liver fibrosis for patients who had not received antiviral treatment remained to be studied. The current study aimed to observe the long-term variation of liver stiffness measurement (LSM), virological and biochemical response on patients without standard antiviral therapy.MethodsA total of 220 patients who were diagnosed with chronic HBV infection, who had not reached the standard of antiviral therapy, and completed a follow-up date of over 2 years in the First Affiliated Hospital of Xi’an Jiaotong University from 2012 to 2018 were retrospectively enrolled. According to the changes of LSM in baseline and follow-up period, the patients were divided into regression group, non-progressive group, and progressive group. The virological and biochemical characteristics of each group were analyzed.ResultsAmong the 220 patients, 153 patients (69.5%) had no progress in LSM degree. Alanine aminotransferase (ALT), HBV DNA, and HBsAg in a few patients increased or slightly decreased, while the vast majority remained in a relatively stable state. 89.5% (137/153) of the non-progressive patients were in grade F0. In addition, 58 patients showed spontaneous improvement with a decreasing rate of 0.460 kPa per year. Patients with ALT of 1-2 ULN had a statistically significant decrease in LSM improvement compared to patients with normal ALT. 82.8% of the LSM-improving patients showed baseline LSM of F1-F3. Only 9 patients showed LSM deterioration, however, which could not be explained by virus replication or necroinflammatory activity. ConclusionsFor patients unsatisfying standard antiviral therapy, most patients with baseline LSM of F0 grade fail to progress, and patients with baseline LSM of F1-F3 show a decrease during follow-up, LSM progression occurs in 4.1% of patients.
Repeated measurement quantitative data is a common data type in clinical studies, and is frequently utilized to assess the therapeutic effects of the intervention measures at a single time point in clinical trials. This study clarifies the concepts and calculation methods for sample size estimation of repeated measurement quantitative data, in order to explore the research question of "comparing group differences at a single time point", from three perspectives: the primary research questions in clinical studies, the main statistical analysis methods and the definitions of the primary outcome indicators. Discrepancies in sample sizes calculated by various methods under different correlation coefficients and varying numbers of repeated measurements were examined. The study revealed that the sample size calculation method based on the mixed-effects model or generalized estimating equations accounts for both the correlation coefficient and the number of repeated measurements, resulting in the smallest estimated sample size. Secondly, the sample size calculation method based on covariance analysis considers the correlation coefficient and produces a smaller estimated sample size than the t-test. The t-test based sample size calculation method requires an appropriate approach to be selected according to the definition of the primary outcome measure. The alignment between the sample size calculation method, the statistical analysis method and the definition of the primary outcome measure is essential to avoid the risk of overestimation or underestimation of the required sample size.
Multi-task learning (MTL) has demonstrated significant advantages in the field of physiological signal measurement. This approach enhances the model's generalization ability by sharing parameters and features between similar tasks, even in data-scarce environments. However, traditional multi-task physiological signal measurement methods face challenges such as feature conflicts between tasks, task imbalance, and excessive model complexity, which limit their application in complex environments. To address these issues, this paper proposes an enhanced multi-scale spatiotemporal network (EMSTN) based on Eulerian video magnification (EVM), super-resolution reconstruction and convolutional multilayer perceptron. First, EVM is introduced in the input stage of the network to amplify subtle color and motion changes in the video, significantly improving the model's ability to capture pulse and respiratory signals. Additionally, a super-resolution reconstruction module is integrated into the network to enhance the image resolution, thereby improving detail capture and increasing the accuracy of facial action unit (AU) tasks. Then, convolutional multilayer perceptron is employed to replace traditional 2D convolutions, improving feature extraction efficiency and flexibility, which significantly boosts the performance of heart rate and respiratory rate measurements. Finally, comprehensive experiments on the Binghamton-Pittsburgh 4D Spontaneous Facial Expression Database (BP4D+) fully validate the effectiveness and superiority of the proposed method in multi-task physiological signal measurement.
To realize the accurate positioning and quantitative volume measurement of tumor in head and neck tumor CT images, we proposed a level set method based on augmented gradient. With the introduction of gradient information in the edge indicator function, our proposed level set model is adaptive to different intensity variation, and achieves accurate tumor segmentation. The segmentation result has been used to calculate tumor volume. In large volume tumor segmentation, the proposed level set method can reduce manual intervention and enhance the segmentation accuracy. Tumor volume calculation results are close to the gold standard. From the experiment results, the augmented gradient based level set method has achieved accurate head and neck tumor segmentation. It can provide useful information to computer aided diagnosis.
Objective
To investigate the application of magnetic resonance imaging (MRI) in preoperative assessment of rectal cancer.
Methods
Combined with the literatures, the MRI features and measurements of rectal tumor staging, extramural vascular invasion, circumferential margin involvement, and the distance between distal margin of the tumor from the anorectal ring and the anal margin were described.
Results
On T2-weighted images (T2WI), T1 staging-tumors were those in which the normal submucosa was replaced by the iso-intensity of tumor tissue without invasion of muscularis propria; T2 staging-tumors were those with extension into the muscularis propria, but not invaded the high-intensity of mesorectal fat; T3 staging-tumors manifested as the rectal tumor penetrated into the muscularis propria and invaded the high-intensity of mesorectal fat; T4 staging-tumors manifested as the tumor invaded adjacent structures or organs. The metastatic lymph nodes were showed with irregular boundaries and mixed signals on T2WI. The tumor signals could be found in the extramural vascular on T1-weighted images (T1WI), accompanied by irregular distortion and expansion of the blood vessels. On T2WI, metastatic lymph nodes, extramural vascular invasion, and the distance between the residual tumor and the low-signal of mesorectal fascia was within 1 mm, indicated the positive circumferential margin. On T2WI, the distal margin of the tumor was located at the junction of hyperintense submucosa and iso-signal of tumor, the tip of the iso-signal puborectal muscle was the apex of the anorectal ring, and the lowest point of the iso-signal external sphincter was the anal margin.
Conclusion
MRI can provide reliable imaging information for preoperative staging, height measurement, and prognosis of rectal cancer, and it is helpful for early diagnosis and treatment of rectal cancer.
Most of the existing near-infrared noninvasive blood glucose detection models focus on the relationship between near-infrared absorbance and blood glucose concentration, but do not consider the impact of human physiological state on blood glucose concentration. In order to improve the performance of prediction model, particle swarm optimization (PSO) algorithm was used to train the structure paramters of back propagation (BP) neural network. Moreover, systolic blood pressure, pulse rate, body temperature and 1 550 nm absorbance were introduced as input variables of blood glucose concentration prediction model, and BP neural network was used as prediction model. In order to solve the problem that traditional BP neural network is easy to fall into local optimization, a hybrid model based on PSO-BP was introduced in this paper. The results showed that the prediction effect of PSO-BP model was better than that of traditional BP neural network. The prediction root mean square error and correlation coefficient of ten-fold cross-validation were 0.95 mmol/L and 0.74, respectively. The Clarke error grid analysis results showed that the proportion of model prediction results falling into region A was 84.39%, and the proportion falling into region B was 15.61%, which met the clinical requirements. The model can quickly measure the blood glucose concentration of the subject, and has relatively high accuracy.
