Pinned Repositories
awesome-gan-for-medical-imaging
Awesome GAN for Medical Imaging
Cardiac_cycle_feature_learning_architecture
Left ventricular ejection fraction (LVEF) is of significant importance for early identification and diagnosis of cardiac disease, but the estimation of LVEF with consistently reliable and high accuracy is still a great challenge due to the high variability of cardiac structures and the complexity of temporal dynamics in cardiac MRI sequences. The widespread methods of LVEF estimation rely on the left ventricular volume. Thus strong prior knowledge is often necessary, which impedes the ease of use of existing methods as clinical tools. In this paper, we propose a cardiac cycle feature learning architecture to achieve an accurate and reliable estimation of LVEF. Unlike the segmentation-based methods, this architecture uses the direct estimation method and does not rely on strong prior knowledge. Experiments on 2900 left ventricle segments of 145 subjects from short axis MR sequences of multiple lengths prove that our proposed method achieves reliable performance (Correlation Coefficient: 0.946; Mean Absolute Error 2.67; Standard Deviation: 3.23). As the first solution to directly estimate LVEF, our proposed method demonstrates great potential in future clinical applications.
covid-chestxray-dataset
We are building an open database of COVID-19 cases with chest X-ray or CT images.
COVID19_imaging_AI_paper_list
COVID-19 imaging-based AI paper collection
Micro8
imlucaslee's Repositories
imlucaslee/awesome-gan-for-medical-imaging
Awesome GAN for Medical Imaging
imlucaslee/Cardiac_cycle_feature_learning_architecture
Left ventricular ejection fraction (LVEF) is of significant importance for early identification and diagnosis of cardiac disease, but the estimation of LVEF with consistently reliable and high accuracy is still a great challenge due to the high variability of cardiac structures and the complexity of temporal dynamics in cardiac MRI sequences. The widespread methods of LVEF estimation rely on the left ventricular volume. Thus strong prior knowledge is often necessary, which impedes the ease of use of existing methods as clinical tools. In this paper, we propose a cardiac cycle feature learning architecture to achieve an accurate and reliable estimation of LVEF. Unlike the segmentation-based methods, this architecture uses the direct estimation method and does not rely on strong prior knowledge. Experiments on 2900 left ventricle segments of 145 subjects from short axis MR sequences of multiple lengths prove that our proposed method achieves reliable performance (Correlation Coefficient: 0.946; Mean Absolute Error 2.67; Standard Deviation: 3.23). As the first solution to directly estimate LVEF, our proposed method demonstrates great potential in future clinical applications.
imlucaslee/covid-chestxray-dataset
We are building an open database of COVID-19 cases with chest X-ray or CT images.
imlucaslee/COVID19_imaging_AI_paper_list
COVID-19 imaging-based AI paper collection
imlucaslee/Micro8