Particle Swarm Algorithm-Based Analysis of Pelvic Dynamic MRI Images in Female Stress Urinary Incontinence.

This work aimed to study the application of pelvic floor dynamic images of magnetic resonance imaging (MRI) based on the particle swarm optimization (PSO) algorithm in female stress urinary incontinence (SUI). 20 SUI female patients were selected as experimental group, and another 20 healthy females were taken as controls. PSO algorithm, K-nearest neighbor (KNN) algorithm, and back propagation neural network (BPNN) algorithm were adopted to construct the evaluation models for comparative analysis, which were then applied to 40 cases of female pelvic floor dynamic MRI images. It was found that the model proposed had relatively high prediction accuracy in both the training set (87.67%) and the test set (88.46%). In contrast to the control group, there were considerable differences in abnormal urethral displacement, urethral length changes, bladder prolapse, and uterine prolapse in experimental patients (P < 0.05). After surgery, the change of urethral inclination angle was evidently reduced (P < 0.05). To sum up, MRI images can be adopted to assess the occurrence of female SUI with abnormal urethral displacement, shortening of urethra length, bladder prolapse, and uterine prolapse. After surgery, the abnormal urethral movement was slightly improved, but there was no obvious impact on bladder prolapse and uterine prolapse.