Recognition of Occluded Goods under Prior Inference Based on Generative Adversarial Network

Mingxuan Cao; Kai Xie; Feng Liu; Bohao Li; Chang Wen; Jianbiao He; Wei Zhang

doi:10.3390/s23063355

Recognition of Occluded Goods under Prior Inference Based on Generative Adversarial Network

Sensors (Basel). 2023 Mar 22;23(6):3355. doi: 10.3390/s23063355.

Authors

Mingxuan Cao^{1

2}, Kai Xie^{1

2

3}, Feng Liu^{1

2}, Bohao Li^{1

2}, Chang Wen^{3

4}, Jianbiao He⁵, Wei Zhang⁵

Affiliations

¹ School of Electronic and Information, Yangtze University, Jingzhou 434023, China.
² National Electrical and Electronic Experimental Teaching Demonstration Center, Yangtze University, Jingzhou 434023, China.
³ Western Research Institute, Yangtze University, Karamay 834000, China.
⁴ School of Computer Science, Yangtze University, Jingzhou 434023, China.
⁵ School of Computer Science and Engineering, Central South University, Changsha 410083, China.

Abstract

Aiming at the recognition of intelligent retail dynamic visual container goods, two problems that lead to low recognition accuracy must be addressed; one is the lack of goods features caused by the occlusion of the hand, and the other is the high similarity of goods. Therefore, this study proposes an approach for occluding goods recognition based on a generative adversarial network combined with prior inference to address the two abovementioned problems. With DarkNet53 as the backbone network, semantic segmentation is used to locate the occluded part in the feature extraction network, and simultaneously, the YOLOX decoupling head is used to obtain the detection frame. Subsequently, a generative adversarial network under prior inference is used to restore and expand the features of the occluded parts, and a multi-scale spatial attention and effective channel attention weighted attention mechanism module is proposed to select fine-grained features of goods. Finally, a metric learning method based on von Mises-Fisher distribution is proposed to increase the class spacing of features to achieve the effect of feature distinction, whilst the distinguished features are utilized to recognize goods at a fine-grained level. The experimental data used in this study were all obtained from the self-made smart retail container dataset, which contains a total of 12 types of goods used for recognition and includes four couples of similar goods. Experimental results reveal that the peak signal-to-noise ratio and structural similarity under improved prior inference are 0.7743 and 0.0183 higher than those of the other models, respectively. Compared with other optimal models, mAP improves the recognition accuracy by 1.2% and the recognition accuracy by 2.82%. This study solves two problems: one is the occlusion caused by hands, and the other is the high similarity of goods, thus meeting the requirements of commodity recognition accuracy in the field of intelligent retail and exhibiting good application prospects.

Keywords: Generative Adversarial Network (GAN); feature expansion; intelligent retail; recognition of occluded goods; semantic inference; vMF distribution.

Abstract

Grants and funding