PENERAPAN SISTEM CHECKOUT BARANG BERBASIS COMPUTER VISION DENGAN METODE MOBILENETV2-SSD

Barcode-based goods checkout systems are currently widely used by large retail stores and grocery stores, barcode-based goods checkout systems are the most reliable transaction machines and have minimal errors when used. Because of this reliability, making barcode-based goods checkout systems tend to have expensive prices. Therefore, a checkout system is needed that has a cheaper price than a barcode system, so the use of computer vision technology is expected to be one of the options that can be chosen. In this study, we will discuss several things related to the application of a computer vision-based goods checkout system. Build the system using the MobileNetV2-SSD method. The purpose and purpose of using this method is to obtain new knowledge or information about model performance and model detection speed when applied to the goods checkout system. In the results of the research that has been done, it was found that the MobileNetV2 SSD method has an mAP (Mean Average Precision) value of 71.41% and has an average detection speed of 42.3 ms (miliseconds). The MobilenetV2-SSD model tends to be accurate
when it only detects one object.