Efficient workflow for automatic segmentation of the right heart based on 2D echocardiography

Area of Science:

• Medical Imaging
• Cardiovascular Imaging
• Image Processing

Background:

• Accurate segmentation of 2D echocardiography images is crucial for diagnosing cardiovascular diseases.
• Existing methods for echocardiography image processing can be time-consuming and lack standardized evaluation metrics.
• Developing an automated and efficient workflow is essential for clinical application.

Purpose of the Study:

• To develop and validate an automated workflow algorithm for processing 2D echocardiography images.
• To compare various image processing techniques for noise reduction, border clarification, and segmentation.
• To introduce a novel metric, the 'global index,' for evaluating the speed and accuracy of image processing workflows.

Main Methods:

• Acquisition of epicardial 2D echocardiography datasets from porcine hearts during surgery.
• Systematic comparison of multiple filtering (e.g., total variation, bilateral), sharpening (linear), and segmentation (e.g., isoline delineation, region-growing) techniques.
• Development and application of the 'global index' metric for workflow performance assessment.

Main Results:

• The optimal workflow identified includes total variation/bilateral filtering, linear sharpening, isoline delineation/region-growing segmentation, and morphological post-processing.
• The 'global index' effectively quantifies the trade-off between processing speed and accuracy (range 0-1).
• The proposed workflow demonstrated efficiency and accuracy in segmenting right heart echocardiography images.

Conclusions:

• An efficient and accurate automated workflow for 2D echocardiography image processing has been established.
• The 'global index' provides a valuable tool for estimating the performance of image pre-processing and segmentation techniques.
• This workflow facilitates precise diagnosis of cardiovascular diseases through improved image analysis.