Hierarchical learning of weights of neural network for performing multiple analyses

Abstract

Systems and methods are provided for performing medical imaging analysis. Input medical imaging data is received for performing a particular one of a plurality of medical imaging analyses. An output that provides a result of the particular medical imaging analysis on the input medical imaging data is generated using a neural network trained to perform the plurality of medical imaging analyses. Theneural network is trained by learning one or more weights associated with the particular medical imaging analysis using one or more weights associated with a different one of the plurality of medicalimaging analyses. The generated output is outputted for performing the particular medical imaging analysis.

Description

[0001] Cross References to Related Applications [0002] This application claims the benefit of U.S. Provisional Application No. 62 / 456,368, filed February 8, 2017, the disclosure of which is hereby incorporated by reference in its entirety. technical field [0003] The present invention relates generally to medical imaging analysis, and more particularly to a hierarchical approach for exploiting commonality in different medical imaging analyses. Background technique [0004] Medical imaging analysis involves extracting information from medical imaging data for performing medical tasks such as landmark detection, anatomy detection, lesion detection, anatomy segmentation, segmentation and characterization, cross-modality image registration (cross-modality image registration), image denoising, etc. Machine learning methods have been widely used to automate medical image analysis. By learning relationships from large databases of annotated training imaging data, machine learn...

AI Technical Summary

Problems solved by technology

[0005] Conventional CNNs developed to perform specific medical imaging analyzes do not take into account the inherent commonalities between different medical imaging analyses.

Images