Auxiliary updating method, system and device

Abstract

The invention discloses an auxiliary updating method, system and device. The method comprises the following steps: receiving misdiagnose sample information; updating the misdiagnose sample informationto a preset database; training a depth learning model based on all sample information in the database to obtain a first network right file of the deep learning model; and finally, sending the first network right file to an ultrasonic instrument for updating the network right file of the ultrasonic instrument, wherein the misdiagnose sample information comprises misdiagnose sample data and correctlabeling results of misdiagnose samples. The auxiliary updating method, system and device disclosed by the invention update a second network right file of the ultrasonic instrument according to the misdiagnose sample information, so that the ultrasonic instrument can judge the misdiagnose sample information accurately by means of the deep learning model established by the second right file, the misjudgement of the ultrasonic instrument on the misdiagnose samples is avoided to a certain extent, and the diagnostic ability of the ultrasonic instrument is improved.

Description

technical field [0001] The present invention relates to the field of medical technology, and more specifically, to an auxiliary update method, system and device. Background technique [0002] With the development of ultrasound technology, more and more ultrasound instruments are used in the medical field to assist doctors in confirming the condition. [0003] An existing method for an ultrasound instrument to assist a doctor in confirming a disease is as follows: the ultrasound instrument constructs a deep learning model with the help of its own stored network weight files; collects sample data; and makes corresponding judgments on the sample data based on the deep learning model for doctors to use according Judging the results and analyzing the condition. [0004] However, the network weight file in an existing ultrasound instrument is immutable, and it is impossible to make correct predictions for all cases. As time goes by, more and more wrong predictions will be found, ...

AI Technical Summary

Problems solved by technology

[0004] However, the network weight file in an existing ultrasound instrument is immutable, and it is impossible to make correct predictions for all cases. As time goes by, more and more wrong predictions will be found, but the same The wrongly predicted samples will always be wrongly predicted under the same model and cannot be corrected, the accumulated wrongly predicted samples are not fully utilized, and the diagnostic ability of ultrasound equipment cannot be improved

Images