System and method for integrating tomography image reconstruction and radiomix using neural networks

摘要

Computed tomography (CT) examination, diagnostics, or other image analysis tasks may be used to integrate supplementary tomographic image reconstructions and radiomixes or to map tomography raw data directly to diagnostic findings within the machine learning framework. Using the above networks and / or algorithms. One or more reconstruction networks are trained to reconstruct tomography images from a training set of CT projection data. One or more radiomix networks are trained to extract features from the tomography images and associated training diagnostic data. The networks / algorithms are integrated and trained into an end-to-end network. Tomography data from an individual, for example, CT projection data, and a set of other relevant information is input to the end-to-end network and based on features extracted by the end-to-end network A potential diagnosis for the individual is calculated. The systems and methods may be applied to CT projection data, MRI data, nuclear medical imaging data, ultrasound signals, optical data, other types of tomography data, or a combination thereof.