Progressive Local Filter Pruning for Image Retrieval Acceleration

Page view(s)
521
Checked on Apr 07, 2024
Progressive Local Filter Pruning for Image Retrieval Acceleration
Title:
Progressive Local Filter Pruning for Image Retrieval Acceleration
Journal Title:
IEEE Transactions on Multimedia
Publication Date:
04 April 2023
Citation:
Wang, X., Zheng, Z., He, Y., Yan, F., Zeng, Z., & Yang, Y. (2023). Progressive Local Filter Pruning for Image Retrieval Acceleration. IEEE Transactions on Multimedia, 1–11. https://doi.org/10.1109/tmm.2023.3256092
Abstract:
Most image retrieval works aim at learning discrim- inative visual features, while little attention is paid to the retrieval efficiency. The speed of feature extraction is key to the real-world system. Therefore, in this paper, we focus on network pruning for image retrieval acceleration. Different from the classification models predicting discrete categories, image retrieval models usually extract continuous features for retrieval, which are more sensitive to network pruning. Such different characteristics of the retrieval and classification models make the traditional pruning method sub-optimal for image retrieval acceleration. Two points are critical for pruning image retrieval models: preserving the local geometry structure of filters and maintaining the model capacity during pruning. In view of the above considerations, we propose a Progressive Local Filter Pruning (PLFP) method. Specifically, we analyze the local geometry of filter distribution in every layer and select redundant filters according to one new criterion that the filter can be replaced locally by other similar filters. Furthermore, to preserve the model capacity of the original model, the proposed method progressively prune the filter by decreasing the scale of filter weights gradually. We evaluate our method on four scene retrieval datasets, i.e., Oxford5K, Oxford105K, Paris6K, and Paris106K, and one person re-identification dataset, i.e., Market-1501. Extensive experiments show that the proposed method (1) preserves the original model capacity while pruning (2) and achieves superior performance to other widely-used pruning methods.
License type:
Publisher Copyright
Funding Info:
This research / project is supported by the SERC - Central Research Fund
Grant Reference no. : SC23/21-1073CI

This paper was supported by China Scholarship Council (No. 201908350025), Fuxiaquan National Independent Innovation Demonstration Zone collaborative platform project (No. 3502ZCQXT2021009), National Natural Science Foundation of Fujian Province (Nos. 2021J011186, 2020J01266), Natural Science Foundation of Xiamen (No. 3502Z20227073), Scientific Research Fund of Fujian Provincial Education Department (No. JAT200486), the University Industry Research Fund of Xiamen (No. 2022CXY0416).
Description:
© 2023 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
ISSN:
1520-9210
1941-0077
Files uploaded:

File Size Format Action
tmmr1-progressive-local-filter-pruning-for-image-representation.pdf 4.33 MB PDF Request a copy