Mining method based on uncertain brain image data

Abstract

The invention provides a mining method based on uncertain brain image data through a method in the field of artificial intelligence. The mining method comprises the following six steps: processing input brain image data to obtain partial anisotropic, average diffusion coefficient, axial diffusion coefficient and radial diffusion coefficient images; using a probabilistic fiber bundle tracking algorithm PICo for anisotropic images to exacte nerve fiber bundles; carrying out image registration on each image with standard Desikan-Killiany template, carrying out corresponding conversion on the nerve fiber bundles; according to the registered images and the nerve fiber bundles, extracting nerve fiber strength, geometric features and diffusion tensor features; designing a data quality evaluation algorithm to evaluate the feature data quality and filter the feature data; designing an algorithm to analyze the filtered data, so as to eliminate the influence of the uncertainty of the data on analysis by processing and analyzing the input brain image, and to mine data in combination with statistical test and a machine learning algorithm in order to detect the required image features.

Description

technical field [0001] The invention relates to the field of artificial intelligence, in particular to a mining method based on uncertain brain image data. Background technique [0002] The human brain can be divided into many functional regions, which are intricately connected and cooperate with each other to complete cognitive tasks. For a long time, researchers lacked effective brain quantification methods. Until recent decades, with the development of medical imaging technology, such as the emergence of magnetic resonance imaging technology, human beings have better means of measuring and quantifying the brain. [0003] In the field of neuroscience, based on brain imaging data, such as MRI images, it is of great clinical value to analyze the human brain network structure and detect disease-related biomarkers. Although medical imaging technology and brain network reconstruction technology are relatively mature, the data obtained through these technologies actually have ...

AI Technical Summary

Problems solved by technology

However, these techniques lack effective means to measure and deal with data uncertainty

Secondly, most of the research objects of existing research methods focus on a single brain network feature, such as nerve fiber strength, and do not support other important features known in the field, such as tensor field diffusion characteristics

Images