LEACS: a learnable and efficient active contour model with space-frequency pooling for medical image segmentation.

Area of Science:

• Medical image analysis
• Computer-aided diagnosis
• Biomedical imaging

Background:

• Accurate segmentation of regions of interest (ROIs) is crucial for disease diagnosis and monitoring.
• Challenges in medical image segmentation include unrefined boundaries, intensity variations, and acquisition limitations.

Purpose of the Study:

• To develop an end-to-end learnable and efficient active contour segmentation model for medical images.
• To enhance the accuracy and efficiency of ROI delineation and segmentation.

Main Methods:

• Proposed an integrated model combining a global convex segmentation (GCS) module with a light-weighted encoder-decoder convolutional segmentation network with a multiscale attention module (ED-MSA).
• Incorporated space-frequency pooling layers in ED-MSA for precise initial contour generation for GCS.
• Utilized depth-wise separable convolutional residual modules in ED-MSA to mitigate model overfitting.

Main Results:

• The integrated ED-MSA and GCS model demonstrated effectiveness in segmenting ROIs across four challenging medical image datasets.
• The proposed method achieved accurate and efficient delineation, addressing limitations of existing techniques.

Conclusions:

• The developed active contour segmentation model offers a robust solution for medical image analysis.
• This approach has the potential to improve diagnostic accuracy and disease monitoring through enhanced image segmentation.