Abstract
Segmenting echocardiographic images is a crucial step in assessing heart function, as clinical indicators can be obtained by precisely delineating the left ventricle. The success of subsequent heart analyses depends entirely on the precision of this segmentation. However, echocardiography is characterized by ambiguity and heavy background noise interference, making accurate segmentation more challenging. Present methods lack efficiency and are prone to mistakenly segmenting some background noise areas, such as the left ventricular area, due to noise disturbance. To address these issues, we introduce P-Mamba, which integrates the Mixture of Experts (MoE) concept for efficient pediatric echocardiographic left ventricular segmentation. Specifically, we utilize the recently proposed ViM layers from the vision mamba to enhance our model's computational and memory efficiency while modeling global dependencies. In the DWT-based (Discrete Wavelet Transform) Perona-Malik Diffusion (PMD) Block, we introduce a block that suppresses noise while preserving the left ventricle's local shape cues. Consequently, our proposed P-Mamba innovatively combines the PMD's noise suppression and local feature extraction capabilities with Mamba's efficient design for global dependency modeling. We conducted segmentation experiments on two pediatric ultrasound datasets and a general ultrasound dataset, namely Echonet-dynamic, and achieved state-of-the-art (SOTA) results. Specifically, on the Pediatric PSAX (8959 images) and Pediatric A4C datasets (6425 images), we achieved Dice scores of 0.922 and 0.906, respectively; on the EchoNet-Dynamic dataset (19882 images), we achieved a Dice score of 0.931. Leveraging the strengths of the P-Mamba block, our model demonstrates superior accuracy and efficiency compared to established models, including vision transformers with quadratic and linear computational complexity.