Small Object Detection Technology Using Multi-Modal Data Based on Deep Learning

Abstract

Recently, Amazon Web Services (AWS) announced a plan to develop a computer vision artificial intelligence service. The field of computer vision has been receiving constant attention like this, and object detection technology is a very important part in this field of computer vision. Although object detection technology is used in many computer vision fields, object detection technology has many limitations in small object detection and night object detection. Various noise factors degrade the image quality, and it is difficult to expect high accuracy when detecting small objects in this environment. In this paper, we develop a technology that can solve these problems by using multi-modal data. In addition, this object detection technology can be used in various fields by designing and developing a lightweight system that can work well in a low-resource environment.