Detection of Renal Calculi in Ultrasound Image Using Meta-Heuristic Support Vector Machine.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Biomedical Signal Processing

Background:

• Renal ultrasound images often contain speckle noise, which hinders accurate medical analysis.
• Detecting renal calculi (kidney stones) in ultrasound images is crucial for diagnosis and treatment.

Purpose of the Study:

• To develop and evaluate a novel method for detecting renal calculi in noisy ultrasound images.
• To improve the accuracy of kidney stone detection by reducing speckle noise and employing advanced classification techniques.

Main Methods:

• An Adaptive Mean Median (AMM) filter was used to reduce speckle noise in ultrasound images.
• Image segmentation was performed using K-Means clustering.
• Gray-Level Co-occurrence Matrix (GLCM) features were extracted for classification.
• A meta-heuristic Support Vector Machine (SVM) classifier, specifically AMM-PSO-SVM, was employed for renal calculi detection.

Main Results:

• The proposed AMM-PSO-SVM method demonstrated superior performance in detecting renal calculi in noisy ultrasound images compared to conventional methods.
• Achieved an accuracy of 98.8% with a False Acceptance Rate (FAR) of 1.8% and a False Rejection Rate (FRR) of 3.3% on a dataset of 250 kidney ultrasound images.

Conclusions:

• The novel AMM-PSO-SVM technique is a promising approach for accurate object detection in medical ultrasound imaging.
• This method effectively addresses the challenge of speckle noise in ultrasound, enhancing the reliability of renal calculi detection.