Please use this identifier to cite or link to this item:
http://hdl.handle.net/1942/32609
Title: | MetaSearch: Incremental Product Search via Deep Meta-Learning | Authors: | WANG, Qi Liu, Xinchen Liu, Wu Liu, An-An Liu, Wenyin Mei, Tao |
Issue Date: | 2020 | Publisher: | IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC | Source: | IEEE TRANSACTIONS ON IMAGE PROCESSING, 29 , p. 7549 -7564 | Abstract: | With the advancement of image processing and computer vision technology, content-based product search is applied in a wide variety of common tasks, such as online shopping, automatic checkout systems, and intelligent logistics. Given a product image as a query, existing product search systems mainly perform the retrieval process using predefined databases with fixed product categories. However, real-world applications often require inserting new categories or updating existing products in the product database. When using existing product search methods, the image feature extraction models must be retrained and database indexes must be rebuilt to accommodate the updated data, and these operations incur high costs for data annotation and training time. To this end, we propose a few-shot incremental product search framework with meta-learning, which requires very few annotated images and has a reasonable training time. In particular, our framework contains a multipooling-based product feature extractor that learns a discriminative representation for each product, and we also design a meta-learning-based feature adapter to guarantee the robustness of the few-shot features. Furthermore, when expanding new categories in batches during a product search, we reconstruct the few-shot features by using an incremental weight combiner to accommodate the incremental search task. Through extensive experiments, we demonstrate that the proposed framework achieves excellent performance for new products while still guaranteeing the high search accuracy of the base categories after gradually expanding new product categories without forgetting. | Notes: | Liu, WY (corresponding author), Guangdong Univ Technol, Dept Comp, Guangzhou 510006, Peoples R China.; Liu, W (corresponding author), JD com, AI Res, Beijing 100101, Peoples R China. wangqi_6414@sina.com; liuxinchen1@jd.com; liuwu1@jd.com; anan0422@gmail.com; liuwy@gdut.edu.cn; tmei@live.com |
Other: | Liu, WY (corresponding author), Guangdong Univ Technol, Dept Comp, Guangzhou 510006, Peoples R China, JD com, AI Res, Beijing 100101, Peoples R China. wangqi_6414@sina.com; liuxinchen1@jd.com; liuwu1@jd.com; anan0422@gmail.com; liuwy@gdut.edu.cn; tmei@live.com | Keywords: | Product search;Few-shot learning;Incremental search;Meta-learning;Multipooling | Document URI: | http://hdl.handle.net/1942/32609 | Link to publication/dataset: | https://ieeexplore.ieee.org/document/9127791 | ISSN: | 1057-7149 | e-ISSN: | 1941-0042 | DOI: | 10.1109/TIP.2020.3004249 | ISI #: | WOS:000553851400023 | Rights: | Copyright 2020 IEEE - All rights reserved. Use of this web site signifies your agreement to the terms and conditions. | Category: | A1 | Type: | Journal Contribution | Validations: | ecoom 2021 |
Appears in Collections: | Research publications |
Show full item record
WEB OF SCIENCETM
Citations
52
checked on Sep 28, 2024
Page view(s)
46
checked on May 30, 2023
Google ScholarTM
Check
Altmetric
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.