In order to solve imperfection of heart rate extraction by method of traditional ballistocardiogram (BCG), this paper proposes an improved method for detecting heart rate by BCG. First, weak cardiac activity signals are acquired in real time by embedded sensors. Local BCG beats are obtained by signal filtering and signal conversion. Second, the heart rate is estimated directly from the BCG beat without the use of a heartbeat template. Compared with other methods, the proposed method has strong advantages in heart rate data accuracy and anti-interference, and it also realizes non-contact online detection. Finally, by analyzing the data of more than 20,000 heart rates of 13 subjects, the average beat error was 0.86% and the coverage was 96.71%. It provides a new way to estimate heart rate for hospital clinical and home care.
In order to improve the efficiency of protein spots detection, a fast detection method based on CUDA was proposed. Firstly, the parallel algorithms of the three most time-consuming parts in the protein spots detection algorithm: image preprocessing, coarse protein point detection and overlapping point segmentation were studied. Then, according to single instruction multiple threads executive model of CUDA to adopted data space strategy of separating two-dimensional (2D) images into blocks, various optimizing measures such as shared memory and 2D texture memory are adopted in this study. The results show that the operative efficiency of this method is obviously improved compared to CPU calculation. As the image size increased, this method makes more improvement in efficiency, such as for the image with the size of 2 048×2 048, the method of CPU needs 5 2641 ms, but the GPU needs only 4 384 ms.
Surface electromyography (sEMG) has been widely used in the study of clinical medicine, rehabilitation medicine, sports, etc., and its endpoints should be detected accurately before analyzing. However, endpoint detection is vulnerable to electrocardiogram (ECG) interference when the sEMG recorders are placed near the heart. In this paper, an endpoint-detection algorithm which is insensitive to ECG interference is proposed. In the algorithm, endpoints of sEMG are detected based on the short-time energy and short-time zero-crossing rates of sEMG. The thresholds of short-time energy and short-time zero-crossing rate are set according to the statistical difference of short-time zero-crossing rate between sEMG and ECG, and the statistical difference of short-time energy between sEMG and the background noise. Experiment results on the sEMG of rectus abdominis muscle demonstrate that the algorithm detects the endpoints of the sEMG with a high accuracy rate of 95.6%.
Objective To explore the methods of early diagnosis of arteriosclerosis obliterans of lower extremity (ASOLE). Methods The related literatures on ASOLE detection means adopted clinically were reviewed, and their advantages and disadvantages were compared.Results Asymptomatic ASOLE could be discovered by determination of ankle brachial index (ABI) and toe brachial index (TBI), which was a good index for arterial function assessment of lower extremity. Pulse wave velocity (PWV) was more vulnerable and less sensitive than ABI, and therefore more suitable for screening of a large sample. ASI was an index to assess arterial structure and function, and it had a good correlation with PWV. Flow-mediated dilation (FMD) was a measurement evaluating the function of endothelial cell; Pulse wave measurement was simple, sensitive, and its result was reliable. Color Doppler ultrasonography could localizate the lesion and determine the degree of stenosis at the same time. Multiple-slice CT angiography (MSCTA) was more accurate than color Doppler ultrasonography, but its inherent shortcomings, such as nephrotoxicity of contrast agent, was still need to be resolved. 3D-contrast enhancement magnetic resonance angiography (CEMRA) had little nephrotoxicity, but a combination of other imaging methods was necessary. Microcirculation detections required high consistency of the measurement environment, but they were simple, sensitive and noninvasive, and therefore could be used for screening of ASO. Conclusion Publicity and education of highrisk groups, and reasonable selection of all kinds of detection means, are helpful to improve the early diagnosis of ASOLE.
It is very important for epilepsy treatment to distinguish epileptic seizure and non-seizure. In this study, an automatic seizure detection algorithm based on dual density dual tree complex wavelet transform (DD-DT CWT) for intracranial electroencephalogram (iEEG) was proposed. The experimental data were collected from 15 719 competition data set up by the National Institutes of Health (NINDS) in Kaggle. The processed database consisted of 55 023 seizure epochs and 501 990 non-seizure epochs. Each epoch was 1 second long and contained 174 sampling points. Firstly, the signal was resampled. Then, DD-DT CWT was used for EEG signal processing. Four kinds of features include wavelet entropy, variance, energy and mean value were extracted from the signal. Finally, these features were sent to least squares-support vector machine (LS-SVM) for learning and classification. The appropriate decomposition level was selected by comparing the experimental results under different wavelet decomposition levels. The experimental results showed that the features selected in this paper were different between seizure and non-seizure. Among the eight patients, the average accuracy of three-level decomposition classification was 91.98%, the sensitivity was 90.15%, and the specificity was 93.81%. The work of this paper shows that our algorithm has excellent performance in the two classification of EEG signals of epileptic patients, and can detect the seizure period automatically and efficiently.
Objective To develop an innovative recognition algorithm that aids physicians in the identification of pulmonary nodules. MethodsPatients with pulmonary nodules who underwent thoracoscopic surgery at the Department of Thoracic Surgery, Affiliated Drum Tower Hospital of Nanjing University Medical School in December 2023, were enrolled in the study. Chest surface exploration data were collected at a rate of 60 frames per second and a resolution of 1 920×1 080. Frame images were saved at regular intervals for subsequent block processing. An algorithm database for lung nodule recognition was developed using the collected data. ResultsA total of 16 patients were enrolled, including 9 males and 7 females, with an average age of (54.9±14.9) years. In the optimized multi-topology convolutional network model, the test results demonstrated an accuracy rate of 94.39% for recognition tasks. Furthermore, the integration of micro-variation amplification technology into the convolutional network model enhanced the accuracy of lung nodule identification to 96.90%. A comprehensive evaluation of the performance of these two models yielded an overall recognition accuracy of 95.59%. Based on these findings, we conclude that the proposed network model is well-suited for the task of lung nodule recognition, with the convolutional network incorporating micro-variation amplification technology exhibiting superior accuracy. Conclusion Compared to traditional methods, our proposed technique significantly enhances the accuracy of lung nodule identification and localization, aiding surgeons in locating lung nodules during thoracoscopic surgery.
ObjectiveTo systematically evaluate the efficacy and safety of computer-aided detection (CADe) and conventional colonoscopy in identifying colorectal adenomas and polyps. MethodsThe PubMed, Embase, Cochrane Library, Web of Science, WanFang Data, VIP, and CNKI databases were electronically searched to collect randomized controlled trials (RCTs) comparing the effectiveness and safety of CADe assisted colonoscopy and conventional colonoscopy in detecting colorectal tumors from 2014 to April 2023. Two reviewers independently screened the literature, extracted data, and evaluated the risk of bias of the included literature. Meta-analysis was performed by RevMan 5.3 software. ResultsA total of 9 RCTs were included, with a total of 6 393 patients. Compared with conventional colonoscopy, the CADe system significantly improved the adenoma detection rate (ADR) (RR=1.22, 95%CI 1.10 to 1.35, P<0.01) and polyp detection rate (PDR) (RR=1.19, 95%CI 1.04 to 1.36, P=0.01). It also reduced the missed diagnosis rate (AMR) of adenomas (RR=0.48, 95%CI 0.34 to 0.67, P<0.01) and the missed diagnosis rate (PMR) of polyps (RR=0.39, 95%CI 0.25 to 0.59, P<0.01). The PDR of proximal polyps significantly increased, while the PDR of ≤5 mm polyps slightly increased, but the PDR of >10mm and pedunculated polyps significantly decreased. The AMR of the cecum, transverse colon, descending colon, and sigmoid colon was significantly reduced. There was no statistically significant difference in the withdrawal time between the two groups. Conclusion The CADe system can increase the detection rate of adenomas and polyps, and reduce the missed diagnosis rate. The detection rate of polyps is related to their location, size, and shape, while the missed diagnosis rate of adenomas is related to their location.
This paper performs a comprehensive study on the computer-aided detection for the medical diagnosis with deep learning. Based on the region convolution neural network and the prior knowledge of target, this algorithm uses the region proposal network, the region of interest pooling strategy, introduces the multi-task loss function: classification loss, bounding box localization loss and object rotation loss, and optimizes it by end-to-end. For medical image it locates the target automatically, and provides the localization result for the next stage task of segmentation. For the detection of left ventricular in echocardiography, proposed additional landmarks such as mitral annulus, endocardial pad and apical position, were used to estimate the left ventricular posture effectively. In order to verify the robustness and effectiveness of the algorithm, the experimental data of ultrasonic and nuclear magnetic resonance images are selected. Experimental results show that the algorithm is fast, accurate and effective.
Objective To observe the effect of BMSCs on the cardiac function in diabetes mellitus (DM) rats through injecting BMSCs into the ventricular wall of the diabetic rats and investigate its mechanism. Methods BMSCs isolated from male SD rats (3-4 months old) were cultured in vitro, and the cells at passage 5 underwent DAPI label ing. Thirty clean grade SD inbred strain male rats weighing about 250 g were randomized into the normal control group (group A), the DM group (group B), and the cell transplantation group (group C). The rats in groups B and C received high fat forage for 4 weeks and the intraperitoneal injection of 30 mg/kg streptozotocin to made the experimental model of type II DM. PBS and DAPI-labeledpassage 5 BMSCs (1 × 105/μL, 160 μL) were injected into the ventricular wall of the rats in groups B and C, respectively. After feeding those rats with high fat forage for another 8 weeks, the apoptosis of myocardial cells was detected by TUNEL, the cardiac function was evaluated with multi-channel physiology recorder, the myocardium APPL1 protein expression was detected by Western blot and immunohistochemistry test, and the NO content was detected by nitrate reductase method. Group C underwent all those tests 16 weeks after taking basic forage. Results In group A, the apoptosis rate was 6.14% ± 0.02%, the AAPL1 level was 2.79 ± 0.32, left ventricular -dP/dt (LV-dP/dt) was (613.27 ± 125.36) mm Hg/s (1 mm Hg=0.133 kPa), the left ventricular end-diastol ic pressure (LVEDP) was (10.06 ± 3.24) mm Hg, and the NO content was (91.54 ± 6.15) nmol/mL. In group B, the apoptosis rate was 45.71% ± 0.04%, the AAPL1 level 1.08 ± 0.24 decreased significantly when compared with group A, the LVdP/ dt was (437.58 ± 117.58) mm Hg/s, the LVEDP was (17.89 ± 2.35) mm Hg, and the NO content was (38.91±8.67) nmol/mL. In group C, the apoptosis rate was 27.43% ± 0.03%, the APPL1 expression level was 2.03 ± 0.22, the LV -dP/dt was (559.38 ± 97.37) mm Hg/ s, the LVEDP was (12.55 ± 2.87) mm Hg, and the NO content was (138.79 ± 7.23) nmol/ mL. For the above mentioned parameters, there was significant difference between group A and group B (P lt; 0.05), and between group B and group C (P lt; 0.05). Conclusion BMSCs transplantation can improve the cardiac function of diabetic rats. Its possible mechanismmay be related to the activation of APPL1 signaling pathway and the increase of NO content.
The existing epilepsy seizure detection algorithms have problems such as overfitting and poor generalization ability due to high reliance on manual labeling of electroencephalogram’s data and data imbalance between seizure and interictal periods. An unsupervised learning detection method for epileptic seizure that jointed graph attention network (GAT) and Transformer framework (GAT-T) was proposed. In this method, channel correlations were adaptively learned by GAT encoder. Temporal information was captured by one-dimensional convolution decoder. Combining outputs of the two mentioned above, predicted values for electroencephalogram were generated. The collective anomaly score was calculated and the detection threshold was determined. The results demonstrated that GAT-T achieved the average performance exceeding 90% (or 99%) with a 0.25 s (or 2 s) time segment length, which could effectively detect epileptic seizures. Moreover, the channel association probability matrix was expected to assist clinicians in the initial screening of the epileptogenic zone, and ablation experiments also reflected the significance of each module in GAT-T. This study may assist clinicians in making more accurate diagnostic and therapeutic decisions for epilepsy patients.