Brain glioma auxiliary labeling method and device based on deep learning

Abstract

The invention provides a brain glioma auxiliary labeling device based on deep learning. Due to the fact that the function of automatic segmentation is achieved through the brain glioma algorithm module, the semi-automation of brain glioma data labeling can be achieved, the advantages of flexibility, convenience and everywhere access of application software of a webpage technology end are fully utilized, and a function of generating data by combining a neural network is combined, so the efficiency of a time-consuming and labor-consuming data labeling process in a data set construction process is improved. The brain glioma auxiliary labeling device is derived from data, and the characteristics of deep learning and feedback data generation are fully utilized; huge consumption caused by manuallabeling is solved; the continuous loop iteration of the brain glioma auxiliary labeling device can be achieved; through a data set training algorithm, and the algorithm is used for assisting the data set construction, so the data annotation becomes more efficient.

Description

technical field [0001] The invention belongs to the fields of computer vision, artificial intelligence technology and medical imaging, and relates to an auxiliary labeling system for MRI images of glioma, in particular to an auxiliary labeling system combining a deep learning brain tumor segmentation algorithm and a web-side data labeling tool. Background technique [0002] In the case of the current machine learning technology and the rapid improvement of computer hardware performance, with the deepening of the application of deep learning in the field of computer vision, deep learning has become more and more auxiliary in different fields. Medical imaging, as an important step and inspection method for diagnosis and treatment in contemporary medicine, is of great significance to the screening, diagnosis and subsequent treatment of diseases. A better combination of learning and medical imaging is a trend. [0003] The rapid development of deep learning algorithms in artifi...

AI Technical Summary

Problems solved by technology

However, for a piece of software, this kind of locally installed client software usually requires a large amount of access data, so its convenience is not enough, so it is urgent to solve the problem of carrying data

[0006] Artificial intelligence algorithms can efficiently solve the problem of tumor segmentation, and efficient and accurate segmentation algorithms require high-quality data sets, and the construction of data sets is completely manual labeling, which is a very time-consuming and energy-consuming task. At the same time, due to the strong professionalism in the field of medical imaging, many labeling tasks must be completed by professional and experienced doctors, so it is inefficient to construct a segmentation data set of glioma only relying on manual labeling

Images