Transfer Learning in Building Neural Network Model Case Study

The transfer learning approach to the development of deep learning (DL) models accelerates the solving of the set task. It also allows solving tasks when large data sets required for extracting non-linear mapping functions are not available. This approach can even reduce the need for large computer resources. This paper contains a case study that shows how easily and quickly it is possible to build a DL model on sparse data using a transfer learning approach. Image processing task was chosen because the development of convolutional neural networks (CNN-Convolutional Neural Network) has led to excellent results in solving the challenges of image classification and object localization and detection. The transfer learning approach enabled simplified and accelerated model development for the selected task. The applied process is formalized into separate distinct procedural steps. A system for the automatic detection of Alzheimer’s dementia through MRI (Magnetic Resonance Imaging) images was the task that needed to be solved. The popular CNN AlexNet architecture was used as a starting point for the development of the new model. A minimally modified model developed by transfer learning achieved an average accuracy of over 88% and an average recall of over 91%. These results are quite comparable to human operators.