A polynomial based algorithm for detection of embolism

Karahoca, Adem; Tunga, M. Alper

doi:10.1007/s00500-014-1240-x

Cited by 11 publications

(8 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specifically, upon examining the classifier weights, it may be seen that there is greater emphasis on unidirectionality. The attained high classification sensitivity and specificity are on par with those reported in prior literature [9], [13], [38], [39]. For instance, Darbellay et al [13] reported embolus classification sensitivity of 95% and associated specificity of 97% on a testing data set comprising 600 emboli and 530 artifacts.…”

Section: Discussionsupporting

confidence: 62%

“…For instance, Darbellay et al [13] reported embolus classification sensitivity of 95% and associated specificity of 97% on a testing data set comprising 600 emboli and 530 artifacts. Using seven classification features, Sombune et al [39] recently reported an average classification sensitivity of 91.5%, average specificity of 90.0%, and average accuracy of 90.5%, outperforming the work by Karahoca and Tunga [38]. Brucher and Russel [9] previously proposed using four features in a decision tree: difference in Doppler shift due to dual-frequency insonation (2 and 2.5 MHz), a measure of expected signal duration, emboli presence in a second depth, and unidirectionality.…”

Section: Discussionmentioning

confidence: 94%

See 1 more Smart Citation

A Time-Frequency Approach for Cerebral Embolic Load Monitoring

Imaduddin

LaRovere

Kussman

et al. 2020

IEEE Trans. Biomed. Eng.

View full text Add to dashboard Cite

To enable reliable cerebral embolic load monitoring from high-intensity transient signals (HITS) recorded with single-channel transcranial Doppler (TCD) ultrasound. Methods: We propose a HITS detection and characterization method using a weighted-frequency Fourier linear combiner that estimates baseline Doppler signal power. An adaptive threshold is determined by examining the Doppler signal power variance about the baseline estimate, and HITS are extracted if their Doppler power exceeds this threshold. As signatures from multiple emboli may be superimposed, we analyze the detected HITS in the time-frequency (TF) domain to segment the signals into individual emboli. A logistic regression classification approach is employed to classify HITS into emboli or artifacts. Data were collected using a commercial TCD device with emboli-detection capabilities from twelve children undergoing mechanical circulatory support or cardiac catheterization. A subset of 696 HITS were reviewed, annotated, and split into training and testing sets for developing and evaluating the HITS classification algorithm. Results: The classifier yielded 98% and 96% sensitivity for 100% specificity on training and testing data, respectively. The TF approach decomposed 38% of candidate embolic signals into two or more embolic events that ultimately account for 69% of the overall embolic counts. Our processing pipeline resulted in highly accurate emboli identification and produced emboli counts that were lower (by a median of 64%) compared to the commercial ultrasound system's estimates. Significance: Using only single-channel, single-frequency Doppler ultrasound, the proposed method enables sensitive detection and segmentation of embolic signatures. Our approach paves the way towards accurate real-time cerebral emboli monitoring.

show abstract

Section: Discussionsupporting

confidence: 62%

Section: Discussionmentioning

confidence: 94%

A Time-Frequency Approach for Cerebral Embolic Load Monitoring

Imaduddin

LaRovere

Kussman

et al. 2020

IEEE Trans. Biomed. Eng.

View full text Add to dashboard Cite

show abstract

“…ANFIS was then used to classify the extracted features as non-ES or ES. The results of this study significantly outperformed the results from the combination of features and algorithms of Karahoca and Tunga [15]. In another prior study, we investigated the use of a deep convolutional neural network (CNN) as a tool for cerebral ES detection [13].…”

Section: A Stroke Screening Methodsmentioning

confidence: 82%

Stroke Screening Feature Selection for Arm Weakness Using a Mobile Application

et al. 2020

View full text Add to dashboard Cite

This work studies the features of a proposed automated stroke self-screening application that utilizes the gyroscope and accelerometer devices in smartphones to determine the possible onset of a stroke by assessing arm muscle weakness. The application requires users to perform two arm movements to evaluate arm weakness and pronation: Curl-up and Raise-up. For the purpose of the study, 68 subjects, consisting of 36 stroke patients with symptoms of arm weakness and 32 healthy subjects, consented to participate. A total of 78 handcrafted features were proposed, 26 of which were extracted from Curl-up and Raise-up for each arm. Then, the differences between corresponding features for each arm were calculated. These features were then tested on 63 combinations of three classical feature selection methods, three feature sets (i.e., Curl-up-only features, Raise-up-only features, and both-exercises combined features) and seven well-known classification methods. The results from ten runs of 10-fold cross-validation showed that Curlup-only features achieved an average sensitivity of 83.3%, significantly higher than those of the Raise-uponly features or both-exercises features. From all possible combinations, the random forest classification based on information gain feature selection from Curl-up-only features achieved the most efficient results for arm-weakness-stroke screening. It achieved an average sensitivity of 94.8%, an average specificity of 75.2%, an average accuracy of 84.1%, and an average area under the receiver operating characteristic curve of 85.0%. Our work proposes a novel accessible method to screen symptoms of arm weakness that may indicate the onset of a stroke using a single mobile device. In the future, we can combine this method with other methods of evaluating facial drooping and slurred speech to create a complete Face, Arm, Speech, Time (FAST) assessment application.

show abstract

“…Furthermore, since the early 2000s, several works have used signal processing and machine learning techniques to detect and classify CE. Signal processing techniques mainly focus on frequency-based approaches using Fourier or wavelet transforms, often combined with machine learning techniques such as support vector machines, random forest, naive Bayes, or fuzzy-rules based classifiers [3]- [6]. Deep learning techniques often use time-frequency representations (TFRs) combined with convolutional neural networks (CNN) [7], [8].…”

Section: Introductionmentioning

confidence: 99%

Deep Embedded Clustering Regularization for Supervised Imbalanced Cerebral Emboli Classification Using Transcranial Doppler Ultrasound

Vindas,

Roux,

Guépié

et al. 2023

2023 31st European Signal Processing Conference (EUSIPCO)

View full text Add to dashboard Cite

High intensity transient signals (HITS) can be detected using transcranial Doppler ultrasound monitoring to help stroke prevention. The various types of HITS are by nature imbalanced: solid emboli are rare events, compared to artifacts or gaseous emboli. Therefore, when training deep models on these data, one have to take into account the class imbalance. In this work, we propose a deep embedded clustering (DEC) based regularization technique, DEC-R to handle imbalanced datasets. Our proposed method decompose a classification model into an encoder and a classifier, with DEC-R applied to the former during supervised training. We validate our method on four synthetic datasets and one HITS datasets. The results show that our approach is robust against imbalanced datasets, allowing to increase the classification performances of different models on both, imbalanced (maximum imbalance ratio of 20) and balanced datasets. We achieve an increase of 4.86% and 1.27% in terms of Matthews correlation coefficient on two imbalanced datasets, a synthetic dataset (of 3D points) and a HITS clinical dataset, respectively. Our code can be found in GitHub.

show abstract

A polynomial based algorithm for detection of embolism

Cited by 11 publications

References 28 publications

A Time-Frequency Approach for Cerebral Embolic Load Monitoring

A Time-Frequency Approach for Cerebral Embolic Load Monitoring

Stroke Screening Feature Selection for Arm Weakness Using a Mobile Application

Deep Embedded Clustering Regularization for Supervised Imbalanced Cerebral Emboli Classification Using Transcranial Doppler Ultrasound

Contact Info

Product

Resources

About