Deep-CodecG*: A Generalized Deep Autoencoder for Robust Segmentation of Left Atrium in Cardiac MRIs: Deep-CodecG* for Robust Segmentation of Left Atrium.

• Autor(a) principal: Rawat, Akhilesh
• Data de Publicação: 2024
• Outros Autores: Kumar, Rajeev
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: INFOCOMP: Jornal de Ciência da Computação
• Texto Completo: https://infocomp.dcc.ufla.br/index.php/infocomp/article/view/2927

The left atrium receives oxygenated blood from pulmonary veins and is a vital organ concerningcongestive heart failure. Several deep learning-based architectures and learning methodologies havebeen proposed for left atrium semantic segmentation. These studies have shown good performance inlearning known datasets. However, generalization remains challenging. In this work, we propose a deepauto-encoder architecture with generalization ability which we call Deep-CodecG*. The proposed modelutilized a CNN-based auto-encoder in which the standard convolution is replaced with a two-convolutionlayer block. This proposed model is generalization enabled with a proper parameterization for (near-)optimal performance. The proposed Deep-CodecG* improves performance on unseen test data, a dicescore of 0.95, which is 6.3% higher than that of a standard auto-encoder. The proposed model gave highersensitivity, specificity, Jaccard, and structural similarity values and lower Hausdorff distance indicatingimprovement over an autoencoder with similar two-convolution layer blocks. Though these quantitativeimprovements seem marginal, they are shown to have a significant impact. The segmented left atriumimages match the ground-truth data very closely. Thus, the proposed Deep-CodecG* architecture for leftatrium segmentation exhibits well-generalized and robust performance over various image datasets