EA-UNet Based Segmentation Method for OCT Image of Uterine Cavity

Optical coherence tomography (OCT) image processing can provide information about the uterine cavity structure, such as endometrial surface roughness, which is important for diagnosis of lesions. The accurate segmentation OCT images a key step processing. We proposed an EA-UNet-based model that uses U-Net network structure with multi-scale attention mechanism to improve accuracy images. E(ECA-C) module introduces convolutional layer combined ECA instead max pool, reduces loss feature information, enables focus on features in region be segmented, and suppresses irrelevant enhance network’s feature-extraction capability learning potential. also introduce A (Attention Gates) model’s by using global contextual information. Our experimental results show EA-UNet ability; furthermore, its MIoU, Sensitivity, Specificity indexes are 0.9379, 0.9457, 0.9908, respectively, indicating effectively has better performance.