A hybrid deep learning model integrating spatial, temporal, and radon features achieved classification accuracies of 94.61% and 95.53% for echocardiographic view recognition.
Does a hybrid deep learning model integrating spatial, temporal, and radon features improve classification accuracy of echocardiographic views in patients undergoing TTE?
A novel three-stream deep learning network integrating spatial, temporal, and radon features achieved high accuracy (~95%) in classifying 24 echocardiographic views, facilitating automated diagnostic workflows.
This study introduces a novel analytics-driven approach for classifying transthoracic echocardiographic (TTE) views by integrating hybrid deep learning, transfer learning, and a three-stream network architecture. The proposed method combines spatial features, temporal dynamics, and radon-transformed representations to enhance model performance and distinguish visually similar echocardiographic views. Trained and evaluated on the EchoIR database, which contains 5,901 videos from 757 patients across 24 view classes, including color Doppler views, this model achieved classification accuracies of 94.61 ± 0.89% and 95.53 ± 0.67% on two benchmark evaluation protocols. By embedding both temporal and Radon-based features, the approach demonstrates significant improvements in predictive accuracy and computational efficiency over existing methods. In the clinical context, accurate view classification is essential for reliable labeling, dataset construction, and the development of automated diagnostic tools. By providing accurate predictions across a large set of echocardiograms, the proposed method facilitates scalable and reliable cardiovascular diagnostic workflows. • Propose a hybrid deep learning model for the accurate classification of cardiac imaging views. • Integrate spatial, temporal, and transformed data to enhance diagnostic image analysis. • Train a three-stream network on a labeled dataset of 5,901 echocardiographic videos. • Achieve precision near expert level using hybrid modeling in cardiac view recognition. • Improve diagnostic accuracy through analytics-driven feature fusion and classification.
Mohammadi et al. (Fri,) conducted a other in Echocardiographic view classification (n=757). Hybrid deep learning model (three-stream network architecture) vs. Existing methods was evaluated on Classification accuracy on benchmark evaluation protocols. A hybrid deep learning model integrating spatial, temporal, and radon features achieved classification accuracies of 94.61% and 95.53% for echocardiographic view recognition.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: