ABSTRACT
@#Objective To propose a heart sound segmentation method based on multi-feature fusion network. Methods Data were obtained from the CinC/PhysioNet 2016 Challenge dataset (a total of 3 153 recordings from 764 patients, about 91.93% of whom were male, with an average age of 30.36 years). Firstly the features were extracted in time domain and time-frequency domain respectively, and reduced redundant features by feature dimensionality reduction. Then, we selected optimal features separately from the two feature spaces that performed best through feature selection. Next, the multi-feature fusion was completed through multi-scale dilated convolution, cooperative fusion, and channel attention mechanism. Finally, the fused features were fed into a bidirectional gated recurrent unit (BiGRU) network to heart sound segmentation results. Results The proposed method achieved precision, recall and F1 score of 96.70%, 96.99%, and 96.84% respectively. Conclusion The multi-feature fusion network proposed in this study has better heart sound segmentation performance, which can provide high-accuracy heart sound segmentation technology support for the design of automatic analysis of heart diseases based on heart sounds.
ABSTRACT
@#Objective To explore the application of Tsetlin Machine (TM) in heart beat classification. Methods TM was used to classify the normal beats, premature ventricular contraction (PVC) and supraventricular premature beats (SPB) in the 2020 data set of China Physiological Signal Challenge. This data set consisted of the single-lead electro-cardiogram data of 10 patients with arrhythmia. One patient with atrial fibrillation was excluded, and finally data of the other 9 patients were included in this study. The classification results were then analyzed. Results The classification results showed that the average recognition accuracy of TM was 84.3%, and the basis of classification could be shown by the bit pattern interpretation diagram. Conclusion TM can explain the classification results when classifying heart beats. The reasonable interpretation of classification results can increase the reliability of the model and facilitate people's review and understanding.
ABSTRACT
@#Objective To recognize the different phases of Korotkoff sounds through deep learning technology, so as to improve the accuracy of blood pressure measurement in different populations. Methods A classification model of the Korotkoff sounds phases was designed, which fused attention mechanism (Attention), residual network (ResNet) and bidirectional long short-term memory (BiLSTM). First, a single Korotkoff sound signal was extracted from the whole Korotkoff sounds signals beat by beat, and each Korotkoff sound signal was converted into a Mel spectrogram. Then, the local feature extraction of Mel spectrogram was processed by using the Attention mechanism and ResNet network, and BiLSTM network was used to deal with the temporal relations between features, and full-connection layer network was applied in reducing the dimension of features. Finally, the classification was completed by SoftMax function. The dataset used in this study was collected from 44 volunteers (24 females, 20 males with an average age of 36 years), and the model performance was verified using 10-fold cross-validation. Results The classification accuracy of the established model for the 5 types of Korotkoff sounds phases was 93.4%, which was higher than that of other models. Conclusion This study proves that the deep learning method can accurately classify Korotkoff sounds phases, which lays a strong technical foundation for the subsequent design of automatic blood pressure measurement methods based on the classification of the Korotkoff sounds phases.
ABSTRACT
OBJECTIVE:To compare th e difference of the components of volatile oil in Citrus medica from different producing areas. METHODS :The volatile oil of C. medica from 10 different producing areas was extracted with steam distillation ,and the yield was calculated. The components of the volatile oil of C. medica from different producing areas were analyzed by GC-MS. The compounds were retrieved from NIST 14.L mass spectrum database and identified. Relative mass fraction of chemical component was determined by peak area normalization method. Cluster analysis of samples were performed by using SPSS 20.0 software. RESULTS:The yields of volatile oil of C. medica from 10 different producing areas were 0.10%-1.75%,among which sample from Qianwei county in Leshan city of Sichuan province was the highest (1.75%). A total of 66 components were identified in the volatile oil of C. medica from different producing areas ,with a relative molecular weight of 126.20-392.66. The majority was C 10 and C 15 compounds;isomers with relative molecular weight of 136,154 took up the great proportion ,which were mainly cycloalkane monoterpenes. There were 12 common components in the volatile oil of C. medica from different areas ,which were limonene(24.90%),terpinene(14.71%),(-)-4 terpineol(2.88%),citral(2.33%),α-myrrhene(2.33%),geraniol(1.52%), α-pinene(1.37%),trans bergamot olene (1.16%),isoterpinene(1.13%),methyl palmitate (1.12%),linalool(1.09%)and geranyl acetate(1.04%)according to relative mass fraction ;8 of them were monoterpenes ,2 were sesquiterpenes and 2 were esters. There were 4 categories of C. medica from different producing areas ,i.e. S 2,S4,S6 clustered into one ;S1,S3,S7,S8 clustered into one ; S5 and S 10 clustered into one ;S9 as one . CONCLUSIONS : There are some difference of the components in volatile oil of medica from different producing areas ,and the content of the same component also has great difference in the volatile oil of C. medica from different producing areas.
ABSTRACT
OBJECTIVE: To optimize the purification technology of total flavonoids from Sparganium stoloniferum. METHODS: Separation and purification by macroporus adsorption resin, using sample solution pH, flow rate and concentration of eluent, the purification rate of total flavonoids as evalution indexes, the purification technology of total flavonoids from S. stoloniferum were optimized by Box-Behnken design-response surface methodology based on single factor test. Validation test was conducted. RESULTS: The optimal purification technology was sample solution pH 4.8, flow rate of eluent 2.0 BV/h, concentration of eluent 72%. The purification rate of total flavonoids in 3 batches of samples was 72.34% (RSD=1.77%, n=3) in validation test, relative errors of which to predicted value (73.99%) was 2.13%. CONCLUSIONS: The optimal purification technology is stable and feasible, and can be used for the purification of total flavonoids from S. stoloniferum.
ABSTRACT
Tinnitus is a common clinical symptom. Researches have shown that fractal sound can effectively treat tinnitus. But current fractal sound is usually synthesized based on constant notes via fractal algorithm, which lead to monotony of synthesized fractal sound. So it is difficult to achieve personalized match. Clinical datas have confirmed that it is common to match tinnitus sound with nature sound and it has a good effect on regulating negative emotion and relieving tinnitus via some natural sound. Therefore, a new method of personalized synthesizing tinnitus rehabilitation sound based on iterative function system (IFS) fractal algorithm is proposed in this paper. This method firstly generates personalized audio library based on natural sound, then tinnitus rehabilitation sound is synthesized via IFS fractal algorithm. Simulation results show that rehabilitation sound in this paper can meet the basic requirements of tinnitus therapy sound and can match tinnitus sound by controlling personalized audio library. So it has reference significance to the treatment of tinnitus sound therapy.
ABSTRACT
Tinnitus is a common clinical symptom and its occurrence rate is high. It seriously affects life quality of the patients. Scientific researches show that listening some similar and none-repetitive music can relieve tinnitus to some extent. The overall music accorded with self-similarity character by the direct mapping method based on chaos. However, there were often the same tones continuous repeating a few times and tone mutations. To solve the problem, this paper proposes a new method for tinnitus rehabilitation sound synthesis based on pentatonic scale, chaos and musical instrument digital interface (MIDI). Experimental results showed that the tinnitus rehabilitation sounds were not only self-similar and incompletely reduplicate, but also no sudden changes. Thus, it has a referential significance for tinnitus treatment.
Subject(s)
Humans , Acoustic Stimulation , Music , Sound , Tinnitus , RehabilitationABSTRACT
Tinnitus is a subjective sensation of sound without external stimulation. It has become ubiquitous and has therefore aroused much attention in recent years. According to the survey, ameliorating tinnitus based on special music and reducing pressure have good effects on the treatment of it. Meantime, vicious cycle chains between tinnitus and bad feelings have been broken. However, tinnitus therapy has been restricted by using looping music. Therefore, a method of generating fractal tones based on musical instrument digital interface (MIDI) technology and pink noise has been proposed in this paper. The experimental results showed that the fractal fragments were self-similar, incompletely reduplicate, and no sudden changes in pitches and would have a referential significance for tinnitus therapy.
Subject(s)
Humans , Acoustic Stimulation , Fractals , Music , Noise , Tinnitus , RehabilitationABSTRACT
Masking therapy can make patients accustom to tinnitus. This therapy is safe and easy to implement, so that it has become a widely used treatment of curing tinnitus. According to surveys of tinnitus sounds, cicada sound is one of the most usual tinnituses. Meanwhile, we have not hitherto found published papers concerning how to synthesize cicada sound and to use it to ameliorate tinnitus. Inspired by the human acoustics theory, we proposed a method to synthesize medical masking sound and to realize the diversity by illustrating the process of synthesizing various cicada sounds. In addition, energy attenuation problem in spectrum shifting process has been successfully solved. Simulation results indicated that the proposed method achieved decent results and would have practical value for the future applications.