Saved in:
Bibliographic Details
Main Authors: Deng, Boyang, Lin, Meiyan, Long, Shoulun
Format: Preprint
Published: 2022
Subjects:
Online Access:https://arxiv.org/abs/2206.05730
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913788909846528
author Deng, Boyang
Lin, Meiyan
Long, Shoulun
author_facet Deng, Boyang
Lin, Meiyan
Long, Shoulun
contents Developing data-efficient instance detection models that can handle rare object categories remains a key challenge in computer vision. However, existing research often overlooks data collection strategies and evaluation metrics tailored to real-world scenarios involving neural networks. In this study, we systematically investigate data collection and augmentation techniques focused on object occlusion, aiming to mimic occlusion relationships observed in practical applications. Surprisingly, we find that even a simple occlusion mechanism is sufficient to achieve strong performance when introducing new object categories. Notably, by adding just 15 images of a new category to a large-scale training dataset containing over half a million images across hundreds of categories, the model achieves 95\% accuracy on an unseen test set with thousands of instances of the new category.
format Preprint
id arxiv_https___arxiv_org_abs_2206_05730
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Adding New Categories in Object Detection Using Few-Shot Copy-Paste
Deng, Boyang
Lin, Meiyan
Long, Shoulun
Computer Vision and Pattern Recognition
Developing data-efficient instance detection models that can handle rare object categories remains a key challenge in computer vision. However, existing research often overlooks data collection strategies and evaluation metrics tailored to real-world scenarios involving neural networks. In this study, we systematically investigate data collection and augmentation techniques focused on object occlusion, aiming to mimic occlusion relationships observed in practical applications. Surprisingly, we find that even a simple occlusion mechanism is sufficient to achieve strong performance when introducing new object categories. Notably, by adding just 15 images of a new category to a large-scale training dataset containing over half a million images across hundreds of categories, the model achieves 95\% accuracy on an unseen test set with thousands of instances of the new category.
title Adding New Categories in Object Detection Using Few-Shot Copy-Paste
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2206.05730