Left Ventricle Full Quantification Using Deep Layer Aggregation Based Multitask Relationship Learning

Fully Automated Quantification of Cardiac Indices from Cine MRI Using a Combination of Convolution Neural Networks

“…From our knowledge, the proposed method achieved the best results for the eight LV parameters (Acav, Amyo, IS, I, IL, AL, A, AS) and similar results for three parameters (dim1, dim2, and dim3) when compared to the state-of-the-art methods [8], [9], [19]. Overall, it exhibited a higher correlation and lower prediction errors for all parameters.…”

“…The DenseNets training took 1 hour and regression inference 8 seconds for all parameters. Table I shows the dice index of the method (the average of five folds and the best fold) for epicardium and endocardium and the results from the best competitors of LVQuan18: Xue et al [8] (reference provided by LVQuan2019), Li et al (1 st place) [9], Kerfoot et al [20] (2 nd place) and Guo et al (3 rd place) [19] .. Xue et al [8], and Li et al [9] methods do not have a segmentation step, thus no comparison is possible for this step of our method. Table II presents the regression results for the 11 parameters, with values of MAE and correlation coefficient (ρ) for both the average of five folds and the best folds.…”

“…Tan et al [8] developed and validated a fully automated neural network regression-based algorithm for segmentation of the LV, with full coverage from apex to base across all cardiac phases, utilizing both SA and long-axis LA scans. Xue et al [9] proposed a deep multitask relationship-learning network to quantify, for every frame in the whole cardiac cycle, multiple types of cardiac indices, such as cavity and myocardium areas, regional wall thicknesses, LV dimension and cardiac phase from SA. Guo et al [10] employed a basic U-net CNN to generate LV segmentation probability maps.…”

“…Guo et al [10] employed a basic U-net CNN to generate LV segmentation probability maps. The derived LV segmentation results were used to quantify the same LV indices presented by Xue et al [9].…”

SK-Unet: An Improved U-Net Model with Selective Kernel for the Segmentation of Multi-sequence Cardiac MR

“…Hence, an automatic segmentation method for this task is desirable in clinical setups. Currently, a majority of studies for the automatic cardiac segmentation are based on the cine CMR sequence [3][4][5][6][7][8], since the cine CMR sequence has the ability to capture the cardiac motions during the whole cardiac cycle and can present clear boundary [1]. The LGE CMR sequence enhances the representation of the infarcted myocardium and is routinely used in the clinical diagnosis of MI.…”