A Classification Method Based on Uniformity of Contrast-enhanced Ultrasound Images

Abstract

The invention discloses a classification method based on the uniformity degree of ultrasound contrast images. The method comprises the following steps: preprocessing an ultrasonic radiography image; obtaining a plurality of rectangular sub-regions of interest, completing classification and marking on each rectangular sub-region of interest, then extracting multiple groups of texture features fromeach rectangular sub-region of interest, screening each group of texture features combining a principal component analysis method to obtain a classification training sample; decomposing the multi-classification problem into a plurality of binary problems. and for each binary classification problem, training by using a classification training sample to obtain a corresponding binary classification QDA model, finally, carrying out classification prediction on unlabeled ultrasound contrast images by using the obtained plurality of binary classification QDA models, and determining a classificationresult of the ultrasound contrast images in a voting manner. Therefore, the learning difficulty of the classification model can be reduced, and the training efficiency of the classification model is improved.

Description

technical field [0001] The invention relates to the technical field of ultra-medical image processing, in particular to a classification method based on the uniformity degree of contrast-enhanced ultrasound images. Background technique [0002] Medical imaging plays an important role in the stage of medical diagnosis, but doctors need a lot of professional knowledge and clinical experience to make judgments through medical images. The judgment results are easily affected by the doctor's subjective factors, and the diagnosis results are unstable. [0003] As a medical image detection method, ultrasonography has the advantages of being non-invasive, painless, and free from the influence of ionizing radiation. Compared with CT and MRI, ultrasound images have higher image resolution, and their real-time dynamic images can better display tumors. Interior details. With the continuous improvement of the resolution of ultrasound instruments and the continuous improvement of the fre...

AI Technical Summary

Problems solved by technology

However, the gray-level co-occurrence matrix requires a large computational overhead and takes up more storage space.

Images