Fusarium Damaged Kernels Detection Using Transfer Learning on Deep Neural Network Architecture

The present work shows the application of transfer learning for a pre-trained deep neural network (DNN), using a small image dataset ($\approx$ 12,000) on a single workstation with enabled NVIDIA GPU card that takes up to 1 hour to complete the training task and archive an overall average accuracy of $94.7\mathrm{\%}$. The DNN presents a $20\mathrm{\%}$ score of misclassification for an external test dataset. The accuracy of the proposed methodology is equivalent to ones using HSI methodology $(81\mathrm{\%}-91\mathrm{\%})$ used for the same task, but with the advantage of being independent on special equipment to classify wheat kernel for FHB symptoms.