FSS-Net: Frequency-Spatial Synergy Network with Wavelet Attention for Carotid Artery Ultrasound Segmentation
Researchers have developed a new deep learning model called FSS-Net for segmenting carotid arteries in ultrasound images. This Frequency-Spatial Synergy Network incorporates wavelet transforms and attention mechanisms to effectively handle challenges like speckle noise and blurred boundaries. Experiments show FSS-Net achieves a high Dice score of 96.46% and demonstrates robustness in low signal-to-noise conditions, indicating potential for clinical applications in identifying abnormal tissue masses. AI
IMPACT New segmentation model offers improved accuracy and robustness for medical ultrasound analysis.