A self-attention-based destruction and construction learning fine-grained image classification method for retail product
- PDF / 901,122 Bytes
- 10 Pages / 595.276 x 790.866 pts Page_size
- 62 Downloads / 184 Views
(0123456789().,-volV)(0123456789(). ,- volV)
S.I. : DEEP LEARNING APPROACHES FOR REALTIME IMAGE SUPER RESOLUTION (DLRSR)
A self-attention-based destruction and construction learning finegrained image classification method for retail product recognition Wenyong Wang1 • Yongcheng Cui1 • Guangshun Li2
•
Chuntao Jiang3 • Song Deng4
Received: 11 January 2020 / Accepted: 17 June 2020 Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract Retail products belonging to the same category usually have extremely similar appearance characteristics such as colors, shapes, and sizes, which cannot be distinguished by conventional classification methods. Currently, the most effective way to solve this problem is fine-grained classification methods, which utilize machine vision ? scene to perform fine feature representations on a target local region, thereby achieving fine-grained classification. Fine-grained classification methods have been widely used for recognizing birds, cars, airplanes, and many others. However, the existing fine-grained classification methods still have some drawbacks. In this paper, we propose an improved fine-grained classification method based on self-attention destruction and construction learning (SADCL) for retail product recognition. Specifically, the proposed method utilizes a self-attention mechanism in the destruction and construction of image information in an end-toend fashion so that to calculate a precise fine-grained classification prediction and large information areas in the reasoning process. We test the proposed method on the Retail Product Checkout (RPC) dataset. Experimental results demonstrate that the proposed method achieved an accuracy above 80% in retail commodity recognition reasoning, which is much higher than the results of other fine-grained classification methods. Keywords Fine-grained classification Multi-attribute recognition Self-attention learning
1 Introduction Currently, hundreds of millions of retail products traffic every day, both online and offline. Commodity retailing is a labor-intensive industry with a very high cost in manual checkouts. With the development of artificial intelligence and deep neural network, it is an irresistible trend to
& Guangshun Li [email protected] 1
College of Information Science and Technology, Northeast Normal University, Changchun, China
2
School of Information Science and Engineering, Qufu Normal University, Rizhao, China
3
School of Mathematics and Big Data, Foshan University, Foshan, China
4
Institute of Advanced Technology, Nanjing University of Post and Telecommunication, Nanjing, China
achieve automatic checkouts of retail products using various image classification methods. Over the past decade, object recognition methods have made steady progress in large-scale data annotation and complex model design. However, there are still a series of problems in recognizing the categories of fine-grained objects (such as birds, butterflies, and vehicles) [1–18]. Therefore, effective fine-grained image c
Data Loading...
