Author

Description

In modern clinical practice, echocardiography is the most vital imaging modality for assessing cardiac function, yet precise segmentation of the Left Ventricle (LV) remains a challenge due to ultrasound artifacts and speckle noise. While standard deep learning architectures excel at segmenting static frames, they often process images in isolation, leading to "flickering" boundaries that lack physical plausibility.

Keywords

chocardiography, Left Ventricle Segmentation, Spatiotemporal Deep Learning, ConvLSTM,