Ultrasound image bladder prolapse automatic grading system

Abstract

The invention relates to an ultrasonic image bladder prolapse automatic grading method. The method includes obtaining an ultrasonic image in a training data set, and performing labeling by a doctor at the same time; performing feature extraction on the obtained ultrasonic image; according to the abovementioned extracted feature values, combined with labeling of the doctor, training a first-layer space-time regression model, forecasting training data to obtain a corresponding displacement diagram, extracting a context feature, and combined with labeling of the doctor, training a second-layer space-time regression model; obtaining an ultrasonic image in a test data set, and performing a labeling forecast by the doctor at the same time; performing feature extraction on the ultrasonic image in the test data set, and utilizing the first-layer space-time regression model to obtain a corresponding diagram; extracting a context feature from the displacement diagram, and utilizing the second-layer space-time regression model to forecast a final displacement diagram and performing identification; and performing automatic grading on bladder prolapse according to a standard. The invention also relates to an ultrasonic image bladder prolapse automatic grading system. The ultrasonic image bladder prolapse automatic grading method and system provided by the invention can reduce an error between groups while improving the examination efficiency.

Description

technical field [0001] The invention relates to an automatic grading system for ultrasonic image bladder prolapse. Background technique [0002] Bladder prolapse is a common female condition in which the bladder bulges out of the vagina due to a defect in the supporting structures of the female pelvic floor. Accurate bladder prolapse grading results are critical for treatment, eg, mild patients do not require treatment, while severe patients may require surgery. [0003] The pelvic organ prolapse quantification system (POP-Q) is widely used in the diagnosis of bladder prolapse. The evaluation system involves many complicated measurement processes, which makes it difficult to implement in clinic. [0004] In recent years, transperineal ultrasonography (TPUS) has been used as an effective method in the classification of bladder prolapse. Its advantages lie in no radiation, relatively comfortable, low cost and real-time imaging capabilities. Generally speaking, the ultrasou...

AI Technical Summary

Problems solved by technology

However, this task is very challenging

First, localization of the pubic symphysis and its inferior border is very difficult, even for experienced physicians, due to the blurry ultrasound images

Second, due to acoustic ray attenuation, speckle noise, and acoustic shadowing, the boundaries of the bladder are missing and inconspicuous, which also makes the task of segmenting the bladder more difficult.

These large differences make the goal of automating the technology more difficult

Images