Logo Detection with Artificial Intelligent
Karn Pinitjitsamut, Kanabadee Srisomboon, Wilaiporn Lee
Abstract
A logo detector is a device that detects illegal produced or counterfeit brand products. The focus of this research is to detect the product logo and consider the likeness of the sample product logo. Building the detector's product logo detector, we used the image detection process with a darknet framework and YOLO algorithm. Through this process, the logo of the copyright products is being set as sample product data for having a dataset. Furthermore, OpenCV image classification by DNN module is used in Python language to read our dataset and work in Windows OS platform, to create a Graphical User Interface (GUI) simply including the creating a function to support the various application. The YOLO algorithm will be the main variable in this research. With this precision, we can detect the fake logo with 97% confidence scores and for the authentic logo with 99% confidence scores.