Ultrasound-Based Kidney Disease Detection: Resampling Cubature Kalman Filter and Physics-Guided VGG Cross-Contextual Network

Limitations of ultrasound-based identification of kidney disease include poor image resolution, operator reliance, and lack of detection of small tissue anomalies. Diagnostic accuracy may be affected by artefacts or noise, or changes in patient anatomy. It may also have difficulty detecting microscopic structural changes or renal disease in its incipiency. The limitations are overcome by the analysis of the Resampling Cubature Kalman Filter and Physics-Guided VGG Cross-Contextual Networks (RCKF-PG-VGG-CCNet) for kidney disease detection from ultrasound images based on the Kidney Ultrasound Image Dataset. The initial use of the Resampling Cubature Kalman Filter and the first use of Physics-Guided VGG Cross-Contextual Networks is to enhance image quality. Through identifying regions through deviations, the Scale-Aware Meet Transformer (SAMT) performs segmentation. While the Synergistic Fibroblast Optimization (SFO) enhances classification performance through optimization of the search for diseased regions, the Physics-Guided VGG Cross-Contextual Networks (PG-VGG-CCNet) classify kidney disease and identify features. Experimental output from a Python-based experiment demonstrates that RCKF-PG-VGG-CCNet is more efficient than other methods with 99.9% efficiency and 99.8% sensitivity and may enhance the accuracy of kidney disease detection.