Modeling radiation therapy planning using hostile generation network

摘要

Techniques for generating a radiation therapy plan and establishing a machine learning model for the generation and optimization of radiation therapy dose data have been disclosed.An exemplary method of generating dose distributions for radiation therapy using a generative model trained in hostile generation networks isReceiving anatomical data for human subjects showing mapping of anatomical regions for radiotherapy, andGeneration modelProcessing anatomical data as inputTo provide dose data as outputGenerating dose data for radiation therapy corresponding to mapping using a trained generation modelIt includes the step of identifying the dose distribution of radiotherapy for radiotherapy of human subjects based on dose data.Another exemplary method for training the generative model involves the use of hostile training including the placement of conditional hostile generation networks to establish the generation model of the hostile generation network and the value of the identification model.Diagram