Suitable Deep Learning Based for High Accuracy Object Detection in Inventory Management: Systematic Literature Review

In the competitive retail industry, maintaining optimal inventory levels is crucial for ensuring smooth business operations. Accurate inventory management goes beyond a mere recording task. The integration of information technology, especially machine learning for inventory management, has become indispensable for optimizing inventory and achieving cost savings. Object detection techniques, have potential in precisely identifying objects and stock-out. However, so many options for object detection models makes the research process more extensive when determining the appropriate model. This study aims to assist researchers in determining the best object detection model to use in the development of current deep learning-based systems, without the need for extensive preliminary research on all available deep learning models using systematic literature review approach. From the research, it is found that the RetinaNet, YOLO, Faster R-CNN, and Mask RCNN are the most suitable choices for inventory management studies, and could reach high accuracy rate above ${9 7 \%}$.