Enhanced Preprocessing and Fast PCA Feature Extraction Algorithm for Ultrasound Images of Ovarian Diseases

Clinical diagnosis of ovarian diseases is mainly based on ultrasonography. Due to the susceptibility of ultrasound images to noise, and the special characteristics and defects of ovarian ultrasound images themselves are prominent, there are some limitations in the existing preprocessing and feature extraction methods. To address the above problems, this paper proposes an enhanced preprocessing and fast principal component analysis (PCA) feature extraction algorithm for ovarian ultrasound images. Combining the unique characteristics of ovarian ultrasound images, the image preprocessing algorithm is designed using a strategy that combines image sharpening using Robert’s operator with linear transformation of contrast. The traditional PCA algorithm is improved to reduce and extract ovarian image features more effectively. The experimental results demonstrate that the proposed method can achieve a generalization rate of 89.79%, AUC of 0.9136 and a duration of 1.98s.