Research

Relationship Matters: Relation Guided Knowledge Transfer for Incremental Learning of Object Detectors

CVPR

Cite Paper

Authors

Published on

06/19/2020

Categories

Computational neuroscience Computer Vision CVPR

Standard deep learning based object detectors suffer from catastrophic forgetting, which results in performance degradation on old classes as new classes are incrementally added. There has been a few recent methods that attempt to address this problem by minimizing the discrepancy between individual object proposal responses for old classes from the original and the updated networks. Different from these methods, we introduce a novel approach that not only focuses on what knowledge to transfer but also how to effectively transfer for minimizing the effect of catastrophic forgetting in incremental learning of object detectors. Towards this, we first propose a proposal selection mechanism using ground truth objects from the new classes and then a relation guided transfer loss function that aims to preserve the relations of selected proposals between the base network and the new network trained on additional classes. Experiments on three standard datasets demonstrate the efficacy of our proposed approach over state-of-the-art methods.

This work was presented at a CVPR workshop in 2020.

Please cite our work using the BibTeX below.

@INPROCEEDINGS {9150833,
author = {K. Ramakrishnan and R. Panda and Q. Fan and J. Henning and A. Oliva and R. Feris},
booktitle = {2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)},
title = {Relationship Matters: Relation Guided Knowledge Transfer for Incremental Learning of Object Detectors},
year = {2020},
volume = {},
issn = {},
pages = {1009-1018},
abstract = {Standard deep learning based object detectors suffer from catastrophic forgetting, which results in performance degradation on old classes as new classes are incrementally added. There has been a few recent methods that attempt to address this problem by minimizing the discrepancy between individual object proposal responses for old classes from the original and the updated networks. Different from these methods, we introduce a novel approach that not only focuses on what knowledge to transfer but also how to effectively transfer for minimizing the effect of catastrophic forgetting in incremental learning of object detectors. Towards this, we first propose a proposal selection mechanism using ground truth objects from the new classes and then a relation guided transfer loss function that aims to preserve the relations of selected proposals between the base network and the new network trained on additional classes. Experiments on three standard datasets demonstrate the efficacy of our proposed approach over state-of-the-art methods.},
keywords = {proposals;detectors;knowledge engineering;object detection;training;task analysis;knowledge transfer},
doi = {10.1109/CVPRW50498.2020.00133},
url = {https://doi-ieeecomputersociety-org.ezproxy.canberra.edu.au/10.1109/CVPRW50498.2020.00133},
publisher = {IEEE Computer Society},
address = {Los Alamitos, CA, USA},
month = {jun}
}