Efficient Parameter Mining and Freezing for Continual Object Detection

Angelo Menezes, Augusto Peterlevitz, Mateus Chinelatto, André C. P.L. F. de Carvalho

2024

Abstract

Continual Object Detection is essential for enabling intelligent agents to interact proactively with humans in real-world settings. While parameter-isolation strategies have been extensively explored in the context of continual learning for classification, they have yet to be fully harnessed for incremental object detection scenarios. Drawing inspiration from prior research that focused on mining individual neuron responses and integrating insights from recent developments in neural pruning, we proposed efficient ways to identify which layers are the most important for a network to maintain the performance of a detector across sequential updates. The presented findings highlight the substantial advantages of layer-level parameter isolation in facilitating incremental learning within object detection models, offering promising avenues for future research and application in real-world scenarios.

Paper Citation

in Harvard Style

Menezes A., Peterlevitz A., Chinelatto M. and C. P.L. F. de Carvalho A. (2024). Efficient Parameter Mining and Freezing for Continual Object Detection. In Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 2: VISAPP; ISBN 978-989-758-679-8, SciTePress, pages 466-474. DOI: 10.5220/0012362300003660

in Bibtex Style

@conference{visapp24,
author={Angelo Menezes and Augusto Peterlevitz and Mateus Chinelatto and André C. P.L. F. de Carvalho},
title={Efficient Parameter Mining and Freezing for Continual Object Detection},
booktitle={Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 2: VISAPP},
year={2024},
pages={466-474},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012362300003660},
isbn={978-989-758-679-8},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 2: VISAPP
TI - Efficient Parameter Mining and Freezing for Continual Object Detection
SN - 978-989-758-679-8
AU - Menezes A.
AU - Peterlevitz A.
AU - Chinelatto M.
AU - C. P.L. F. de Carvalho A.
PY - 2024
SP - 466
EP - 474
DO - 10.5220/0012362300003660
PB - SciTePress