The Training Of An Artificial Neural Network For A Detection Task In Medical Images
S. Garg, Carey E. Floyd · 2005
In this paper, an artiikial neural network is constructed to perform a feature detection task in medical images. The neural network model is a Perceptmn and uses the Back Propagation Method learning algorithm with the Generalized Delta Rule to modify the connection strengths. The adaptive weight matrices that result following the training stage of the neural network are a function of which images are presented and a~ shown to correlate to actual retinal cell organization. The development of this organization is signifEant, in light of the fact that nodule detection is a visual processing hsk.