Single-fusion detector: Towards faster multi-scale object detection
Despite recent improvements, the arbitrary sizes of objects still impede the predictive ability of object detectors. Recent solutions combine feature maps of different receptive fields to detect multi-scale objects. However, these methods have large computational costs resulting to slower inference...
Saved in:
Main Authors: | , , , |
---|---|
Format: | text |
Published: |
Animo Repository
2019
|
Subjects: | |
Online Access: | https://animorepository.dlsu.edu.ph/faculty_research/3022 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | De La Salle University |
id |
oai:animorepository.dlsu.edu.ph:faculty_research-4021 |
---|---|
record_format |
eprints |
spelling |
oai:animorepository.dlsu.edu.ph:faculty_research-40212021-11-19T08:27:36Z Single-fusion detector: Towards faster multi-scale object detection Antioquia, Arren Matthew C. Tan, Daniel Stanley Azcarraga, Arnulfo Hua, Kai Lung Despite recent improvements, the arbitrary sizes of objects still impede the predictive ability of object detectors. Recent solutions combine feature maps of different receptive fields to detect multi-scale objects. However, these methods have large computational costs resulting to slower inference time, which is not practical for real-time applications. Contrarily, fusion methods depending on large networks with many skip connections demand larger memory requirement, prohibiting usage in devices with limited memory. In this paper, we propose a more computationally efficient fusion method which integrates higher-order information to low-level feature maps using a single operation. Our method can flexibly adapt to any base network, allowing tailored performance for different computational requirements. Our approach achieves 81.7% mAP at 41 FPS on the PASCAL VOC dataset using ResNet-50 as the base network, which is superior in terms of both speed and mAP as compared to several state-of-the-art baselines, even those which use larger base networks. © 2019 IEEE. 2019-09-01T07:00:00Z text https://animorepository.dlsu.edu.ph/faculty_research/3022 Faculty Research Work Animo Repository Image converters Neural networks (Computer science) Computer Sciences |
institution |
De La Salle University |
building |
De La Salle University Library |
continent |
Asia |
country |
Philippines Philippines |
content_provider |
De La Salle University Library |
collection |
DLSU Institutional Repository |
topic |
Image converters Neural networks (Computer science) Computer Sciences |
spellingShingle |
Image converters Neural networks (Computer science) Computer Sciences Antioquia, Arren Matthew C. Tan, Daniel Stanley Azcarraga, Arnulfo Hua, Kai Lung Single-fusion detector: Towards faster multi-scale object detection |
description |
Despite recent improvements, the arbitrary sizes of objects still impede the predictive ability of object detectors. Recent solutions combine feature maps of different receptive fields to detect multi-scale objects. However, these methods have large computational costs resulting to slower inference time, which is not practical for real-time applications. Contrarily, fusion methods depending on large networks with many skip connections demand larger memory requirement, prohibiting usage in devices with limited memory. In this paper, we propose a more computationally efficient fusion method which integrates higher-order information to low-level feature maps using a single operation. Our method can flexibly adapt to any base network, allowing tailored performance for different computational requirements. Our approach achieves 81.7% mAP at 41 FPS on the PASCAL VOC dataset using ResNet-50 as the base network, which is superior in terms of both speed and mAP as compared to several state-of-the-art baselines, even those which use larger base networks. © 2019 IEEE. |
format |
text |
author |
Antioquia, Arren Matthew C. Tan, Daniel Stanley Azcarraga, Arnulfo Hua, Kai Lung |
author_facet |
Antioquia, Arren Matthew C. Tan, Daniel Stanley Azcarraga, Arnulfo Hua, Kai Lung |
author_sort |
Antioquia, Arren Matthew C. |
title |
Single-fusion detector: Towards faster multi-scale object detection |
title_short |
Single-fusion detector: Towards faster multi-scale object detection |
title_full |
Single-fusion detector: Towards faster multi-scale object detection |
title_fullStr |
Single-fusion detector: Towards faster multi-scale object detection |
title_full_unstemmed |
Single-fusion detector: Towards faster multi-scale object detection |
title_sort |
single-fusion detector: towards faster multi-scale object detection |
publisher |
Animo Repository |
publishDate |
2019 |
url |
https://animorepository.dlsu.edu.ph/faculty_research/3022 |
_version_ |
1718383322937163776 |