Logo Detection Using You Only Look Once (YOLO) Method
This research aimed to identify the logo on a document in the form of a college diploma by utilizing the You Only Look Once algorithm version 4 (YOLOv4) by implementing the Darknet Framework.
Abstract
The growing amount of documents necessitates automatic document classification since manual classification takes longer. Logos allow for defining the source of a document rapidly and accurately. Logos are part of the identity attached to companies, organizations, institutions, and individuals. The placement of the logo on each document varies, as does the image quality of a document. This research aimed to identify the logo on a document in the form of a college diploma. This study utilized diplomas from fifteen universities. This system utilizes the You Only Look Once algorithm version 4 (YOLOv4) by implementing the Darknet Framework. The ultimate YOLOv4 model's results indicated that the system could detect and recognize logos with a mean Average Precision (mAP) of 93.73%.