International Journal of Scientific and Technological Research www.iiste.org ISSN 2422-8702 (Online) Vol 4, No.10, 2018 522 | Page www.iiste.org Product Identification System Design Based on Deep Learning Muhammet Rasit Cesur (Corresponding author) Department of Industrial Engineering, Sakarya University, Turkey E-mail: rcesur@sakarya.edu.tr Elif Cesur Department of Industrial Engineering, Istanbul Medeniyet University, Turkey E-mail: elif.karakaya@medeniyet.edu.tr Ismail Hakki Cedimoglu Department of Industrial Engineering, Sakarya University, Turkey E-mail: cedim@sakarya.edu.tr Orhan Torkul Department of Industrial Engineering, Yalova University, Turkey E-mail: torkul@yalova.edu.tr Alper Okuyan Toyota Boshoku Türkiye E-mail: alper.okuyan@toyota-boshokutr.com The research is financed by Toyota Boshoku, Turkey Abstract Product identification approach has become a crucial issue for the inventory monitoring process. Because of the enhanced product spectrum and many shipment points, it is quite likely to send incorrect orders to the customer erroneously. With the help of this proposed model, particular troubles during product shipment could be solved. In this study, it is aimed to not only delivery right product to the right consumers but also ensures Kanban system inside the facility by performing product identification procedure. Within the scope of this study, a technology - industry integrated system has been designed by using convolutional neural network architecture. Keywords: Convolutional Neural Network, Deep Learning, Product Identification 1. Introduction Recent developments in the field of computer hardware have resulted in huge progress in Artificial Intelligence (AI) technology. The reason behind this fact that complex and complicated problems could be handled within a short amount of time with the help of high computer technology. Thus, the AI method not only could be more efficient to make comprehensively analyses and find more accurate outcomes but also has gained a novel structure to perform dynamically. Conventional AI methods take into whole