A neural network regression model for relative dose computation

This paper describes a neural network (NN) regression model for relative do se computation. The input signal of the the NN model includes depths and fi eld sizes and the output is the relative dose, for example, percentage dept h dose (%DD) or tissue-air ratio (TAX) in this paper. After a functional li nk has been created, the expressing ability of input patterns and the resol ution of neural networks are enhanced. The trained neural network exhibits a good generalization and interpolation ability, and it can be easily exten ded to other relative dose regressions. We present two examples to verify t he fitness of this model. The first calculates the percentage depth dose of a 14 MV x-ray beam for field sizes of 5 cm x 5 cm to 28 cm x 28 cm. The av erage error computed from the NN is less than 0.47% comparing with the orig inal measured %DD. The second example calculates the TAR of a 4 MV x-ray be am for field sizes of 5 cm x 5 cm to 20 cm x 20 cm. In this example, the tr aining data show that the average error computed from the NN is less than 0 .48%, whereas that from the testing data is less than 0.37% after training. Such an NN model can be generalized to fit data for treatment planning wit h any linear accelerator and can also fit data stored as tissue-maximum rat io (TMR) and tissue-phantom ratio (TPR).