Barcode Detection and Classification using SSD (Single Shot Multibox Detector) Deep Learning Algorithm

Abstract

Barcodes have become ubiquitous in this era, as they are widely used in many applications. With the rapid usage of barcodes, thus fast and accurate detecting is essential. We introduce deep learning detector based on SSD (Single Shot Multibox Detector) which is expected to be fast and accurate. This system is expected to be capable of detecting both the barcodes i.e. 1D and 2D barcodes. The detector can also find the rotation of both the barcodes using angle prediction network. The proposed approach is one of the deep learning algorithm which will use a single convolutional neural network and expected to perform well as compared to other detectors such as YOLO (You Only Look Once) in both speed and accuracy.

Keywords: SSD (Single Shot Multibox Detector), angle prediction network, single Convolutional neural network, YOLO (You Only Look Once)