Image-based classification of bladder state using electrical impedance tomography.

摘要: Objective In this study, we examine the potential of using machine learning classification to determine bladder state ('not full', 'full') with electrical impedance tomography (EIT) images pelvic region. Accurate these states would enable urinary incontinence (UI) monitoring alert patient, before involuntary voiding occurs, in a low-cost and discrete manner. Approach Using both numerical experimental data, form datasets that contain diverse observations varying clinical parameters such as volume, urine conductivity, reference used for time-difference imaging. We then classify pixel-wise feature extraction-based techniques. employ principal component analysis, wavelets, image segmentation help create features. Main results The performance was compared across several classifier algorithms. minimum accuracy 77.50%. highest observed 100%, found by combining analysis Gaussian radial based function kernel support vector machine. This combination also offered best trade-off between costs training time memory space. biggest challenge is classifying volumes near separation volume not full full, which choosing most suitable can minimize error. Significance performed first EIT images, achieving high accuracies data. work highlights image-based an device those suffering from UI.