Neural network prediction modeling based on elastographic textural features for meat quality evaluation

Multilayer feedforward neural networks are proposed to capture the nonlinearity between the system inputs and outputs to predict meat quality with the wavelet textural features from the ultrasonic elastograms. This article investigated the efficiency of the training processes and the generalization of the networks using the gradient descent and Levenberg-Marquardt algorithms in backpropagation. It was found that for this application in the case of low epoch training (below several thousand epochs) using the gradient descent algorithm, the Levenberg-Marquardt algorithm was less efficient, and in the case of high epoch training (above several thousand epochs) using the gradient descent algorithm, the Levenberg-Marquardt algorithm was more efficient. In the case of difficult convergence in the gradient descent algorithm, the Levenberg-Marquardt algorithm converged effectively. In either case, the Levenberg-Marquardt algorithm better modeled output variation accounting and network generalization. Weight-decay was further used in the Levenberg-Marquardt backpropagation to improve the generalization of the network models. The leave-one-out procedure was built into every training process to ensure sufficient modeling on limited samples.

Neural network prediction modeling based on elastographic textural features for meat quality evaluation Academic Article