Key result
A developed U-Net model achieved a dice score of 0.94 for segmenting the left atrium, demonstrating higher accuracy than hybridized CNN and LSTM models.
Why the study?
Existing deep learning models, including hybridized CNN and LSTM architectures, showed suboptimal dice scores for left atrium segmentation.
Answers are grounded in this paper and cite their sources.
A modified U-Net deep learning model achieved high accuracy (Dice score 0.94) for left atrium segmentation, outperforming standard hybrid CNN-LSTM models.
Atrium segmentation is a very important task for cardiologists, especially in the left atrium (LA). Since the chances of narrowing are common in that area where blood stops flowing and which causes a heart attack. To tackle this problem, Artificial Intelligence can play an important role in segmenting the blockage area. There are many models for segmenting the left atrium such as Convolutional Neural Network (CNN), bidirectional Long Short Term Memory (LSTM) etc. Hybridize method can be an efficient method for segmenting the left atrium The most common method used is to hybridize CNN and LSTM. By analyzing the above models, it was found that the dice score was not working well even after making the model hybridized. The dice score is considered in the range from 0.85 to 0.92 for segmenting the left atrium. In our work, a U-Net model has been developed for segmenting the left atrium and the model has been developed with a change in parameters. It was observed that this model achieved a dice score of 0.94, which was found comparatively more accurate than any other hybridized model.
No takes yet. Share an insight, caveat, or question.
Aryan et al. (2022) studied Left atrium segmentation. U-Net model with modified parameters vs. Hybridized CNN and LSTM models was evaluated on Dice score for segmenting the left atrium. A developed U-Net model achieved a dice score of 0.94 for segmenting the left atrium, demonstrating higher accuracy than hybridized CNN and LSTM models.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: