Research Results from University of Aizu Update Knowledge of Diagnostics (Bidirectional Copy-Paste Mamba for Enhanced Semi-Supervised Segmentation of Transvaginal Uterine Ultrasound Images).

Researchers from the University of Aizu have developed a semi-supervised model called Bidirectional Copy-Paste Mamba (BCP-Mamba) for segmenting the parametrium region of transvaginal uterine ultrasound images. The model incorporates a U-shaped structure model with a visual state space module and was trained and evaluated using a dataset of 1940 ultrasound images. Comparative analysis showed that the BCP-Mamba model outperformed two other semi-supervised models, achieving a Dice coefficient of 86.55%. The researchers believe that this model can alleviate the workload of experts and improve the accuracy of prediction and diagnosis of uterine-related conditions. [Extracted from the article]