Abstract: Deep learning typically requires a large amount of data and it is computationally expensive for training an application from scratch. ImageNet database has millions of images pertaining to different categories that are acquired by years of hard work. Getting such a database for every application is tough and time consuming. Transfer learning is an alternative to conventional training. Transfer learning results in much faster and easier training of a network. This research set out to evaluate the effect of transfer learning on the performance of a Deep Neural Network (DNN). Pre- trained AlexNet was selected, modified and retrained for 3 image classification applications (gears, connectors and coins) with a modest database. This approach gave 99% classification accuracy using transfer learning. To test the robustness of the network, unknown images were added to one of the classes and the accuracy was reinforced using a probability threshold. This approach succeeded in compensating for the effect of unknowns in the accuracy (The above paper presented in 20th international conference for Engineering Applications in Neural Networks (EANN 2019). The speaker has received an international travel award to present the paper at the conference venue in Greece)

About the Speaker: Keyur D. Joshi holds a PhD in the field of machine vision from Queen’s University. His fields of interests are machine vision, machine learning, artificial intelligence, industrial automation and control systems.