Multi-form medical image data labeling method and system

Abstract

The invention relates to a multi-form medical image data labeling method and system, and the method comprises the steps: receiving a picture uploaded by a user, and checking whether the format is correct or not according to a preset picture format; if yes, uploading the picture, and if not, reminding the user to upload the picture again; extracting picture information to be matched with an information base, and judging whether the picture has annotation or attribute information or not; setting the permission of labeling the picture; providing a labeling tool box for the user according to the disease type; receiving annotation information of the user on the picture, and uploading and storing the annotation information; extracting annotation information of users of the same level for the picture, comparing the annotation information, judging whether the annotation information of the users of the same level is different or not, if not, generating a final annotation result, and if yes, submitting the final annotation result to a next-level user for further annotation. According to the method and system, different labeling tools can be used for labeling images, the labeling requirementsof medical images in various forms are met, and the accuracy of image labeling is improved by adopting a hierarchical labeling process.

Description

technical field [0001] The invention relates to the field of medical artificial intelligence, in particular to a method and system for labeling medical image data in various forms. Background technique [0002] With the maturity of artificial intelligence technology, the scope of application is more extensive, but in the field of medical AI, compared with the rapid development of other fields, it seems to be relatively slow. The main reasons are small data size, poor annotation quality, and the need for cross-field and interdisciplinary professionals for annotation. Collecting, sorting, segmenting and labeling high-quality medical image data, and transforming them into structured data that can be recognized by computers and further processed is the basis of artificial intelligence algorithms. However, at this stage, the medical image data labeling platform cannot meet the labeling requirements of various forms of medical images. On the one hand, different types of diseases ...

AI Technical Summary

Problems solved by technology

[0004] (1) Image labeling at this stage relies on a large number of professional doctors, and doctors need to spend energy on labeling after work

[0005] (2) At this stage, the labeling platform is only used as a tool, and has not simulated a complete labeling and inspection process. Usually, the labeling and structuring of medical images requires a large professional team. At this stage, the labeling tool does not play a team role the role of

[0006] (3) The current labeling tools have poor scalability and high cost of use, and medical images with multiple diseases and problems require a large number of proprietary labeling tools, such as scoliosis, which requires a custom tool to measure the Cobb angle

It is expensive to use existing annotation platforms to customize a large number of annotation tools

Images