Journal: Sensors (Basel, Switzerland)

Article Title: A Crowdsourced AI Framework for Atrial Fibrillation Detection in Apple Watch and Kardia Mobile ECGs

doi: 10.3390/s24175708

Figure Lengend Snippet: The ECG datasets used in this study.

Article Snippet: Unlike other datasets that typically consist of unique ECG recordings from individual subjects, the AliveCor DS3 dataset includes multiple recordings from each individual.

Techniques:

Journal: Sensors (Basel, Switzerland)

Article Title: A Crowdsourced AI Framework for Atrial Fibrillation Detection in Apple Watch and Kardia Mobile ECGs

doi: 10.3390/s24175708

Figure Lengend Snippet: Results of the voting algorithm vs. standalone Apple and Kardia algorithms on different ECG datasets.

Article Snippet: Unlike other datasets that typically consist of unique ECG recordings from individual subjects, the AliveCor DS3 dataset includes multiple recordings from each individual.

Techniques:

Journal: Sensors (Basel, Switzerland)

Article Title: A Crowdsourced AI Framework for Atrial Fibrillation Detection in Apple Watch and Kardia Mobile ECGs

doi: 10.3390/s24175708

Figure Lengend Snippet: The vectorized confusion matrices of the results of the classification of the voting algorithm and standalone Apple and Kardia algorithms on different ECG datasets.

Article Snippet: Unlike other datasets that typically consist of unique ECG recordings from individual subjects, the AliveCor DS3 dataset includes multiple recordings from each individual.

Techniques:

Journal: Sensors (Basel, Switzerland)

Article Title: A Crowdsourced AI Framework for Atrial Fibrillation Detection in Apple Watch and Kardia Mobile ECGs

doi: 10.3390/s24175708

Figure Lengend Snippet: The overall performance of the voting algorithm and each individual base algorithm across multiple ECG datasets.

Article Snippet: Unlike other datasets that typically consist of unique ECG recordings from individual subjects, the AliveCor DS3 dataset includes multiple recordings from each individual.

Techniques:

Journal: Sensors (Basel, Switzerland)

Article Title: A Crowdsourced AI Framework for Atrial Fibrillation Detection in Apple Watch and Kardia Mobile ECGs

doi: 10.3390/s24175708

Figure Lengend Snippet: A performance comparison of the voting algorithm against six base algorithms and proprietary solutions across different datasets. (Datta et al. [ , ], Gliner et al. [ , ], Kropf et al. [ , ], Baydoun et al., Zabihi et al. , and Soliński et al. ). Left column : ROC curves. Right column : precision–recall curves. Upper panels : results on Apple Watch DS dataset. Middle panels : results on AliveCor DS2 dataset. Lower panels : results on AliveCor DS3 dataset. The voting algorithms’ performance (orange lines) is compared with that of the Apple/Kardia algorithms (black x) and six base algorithms (blue shapes). In all panels, the operating point of the voting algorithm is indicated by an orange circle. The consistently superior performance of the voting algorithm is demonstrated by its ROC and precision–recall curves encompassing those of the base algorithms across all datasets, showcasing its robustness and generalizability.

Article Snippet: Unlike other datasets that typically consist of unique ECG recordings from individual subjects, the AliveCor DS3 dataset includes multiple recordings from each individual.

Techniques: Comparison