Mask R-CNN with New Data Augmentation Features for Smart Detection of Retail Products

Human–computer interactions (HCIs) use computer technology to manage the interfaces between users and computers. Object detection systems that convolutional neural networks (CNNs) have been repeatedly improved. Computer vision is also widely applied multiple specialties. However, self-checkouts operating with a faster region-based network (faster R-CNN) image system still feature overlapping cannot distinguish color of objects, so inhibited. This study uses mask R-CNN data augmentation (DA) discrete wavelet transform (DWT) in lieu prevent trivial details images from hindering extraction for deep learning (DL). The experiment results show proposed algorithm allows more accurate efficient similarly colored objects than ResNet 101, but excellent resolution real-time processing smart retail stores.