Few-Shot Object Detection via Sample Processing

Few-shot object detection (FSOD) eliminates the dependence on tremendous instances with manual annotations in conventional object detection. We deem that the scarcity of positive samples is the main reason that restricts the performance of FSOD detectors. In this paper, a novel FSOD model via sample...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Honghui Xu, Xinqing Wang, Faming Shao, Baoguo Duan, Peng Zhang
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2021
Materias:	Few-shot learning image processing machine learning object detection Electrical engineering. Electronics. Nuclear engineering TK1-9971
Acceso en línea:	https://doaj.org/article/73554ba2e8104e98bad6e1cc5b4a79f8
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:73554ba2e8104e98bad6e1cc5b4a79f8
record_format	dspace
spelling	oai:doaj.org-article:73554ba2e8104e98bad6e1cc5b4a79f82021-11-19T00:05:52ZFew-Shot Object Detection via Sample Processing2169-353610.1109/ACCESS.2021.3059446https://doaj.org/article/73554ba2e8104e98bad6e1cc5b4a79f82021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9354609/https://doaj.org/toc/2169-3536Few-shot object detection (FSOD) eliminates the dependence on tremendous instances with manual annotations in conventional object detection. We deem that the scarcity of positive samples is the main reason that restricts the performance of FSOD detectors. In this paper, a novel FSOD model via sample processing, namely, FSSP, is proposed to detect objects accurately with only a few annotated samples, which is based on the structural design of the Siamese network and uses YOLOv3-SPP as the baseline. Central to FSSP are our designed self-attention (SAM) and positive-sample augmentation (PSA) modules. The former attempts to better extract the representative features of hard samples, and the latter expands the number and enriches the scale distribution of positive samples, inhibiting the growth of negative samples. For the fine-tuning phase, we modify the classification loss function to increase the punishment for hard samples. Experiments conducted on the PASCAL VOC and MS COCO datasets confirm that the proposed FSSP achieves competitive detection performance compared with state-of-the-art detectors.Honghui XuXinqing WangFaming ShaoBaoguo DuanPeng ZhangIEEEarticleFew-shot learningimage processingmachine learningobject detectionElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 9, Pp 29207-29221 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Few-shot learning image processing machine learning object detection Electrical engineering. Electronics. Nuclear engineering TK1-9971
spellingShingle	Few-shot learning image processing machine learning object detection Electrical engineering. Electronics. Nuclear engineering TK1-9971 Honghui Xu Xinqing Wang Faming Shao Baoguo Duan Peng Zhang Few-Shot Object Detection via Sample Processing
description	Few-shot object detection (FSOD) eliminates the dependence on tremendous instances with manual annotations in conventional object detection. We deem that the scarcity of positive samples is the main reason that restricts the performance of FSOD detectors. In this paper, a novel FSOD model via sample processing, namely, FSSP, is proposed to detect objects accurately with only a few annotated samples, which is based on the structural design of the Siamese network and uses YOLOv3-SPP as the baseline. Central to FSSP are our designed self-attention (SAM) and positive-sample augmentation (PSA) modules. The former attempts to better extract the representative features of hard samples, and the latter expands the number and enriches the scale distribution of positive samples, inhibiting the growth of negative samples. For the fine-tuning phase, we modify the classification loss function to increase the punishment for hard samples. Experiments conducted on the PASCAL VOC and MS COCO datasets confirm that the proposed FSSP achieves competitive detection performance compared with state-of-the-art detectors.
format	article
author	Honghui Xu Xinqing Wang Faming Shao Baoguo Duan Peng Zhang
author_facet	Honghui Xu Xinqing Wang Faming Shao Baoguo Duan Peng Zhang
author_sort	Honghui Xu
title	Few-Shot Object Detection via Sample Processing
title_short	Few-Shot Object Detection via Sample Processing
title_full	Few-Shot Object Detection via Sample Processing
title_fullStr	Few-Shot Object Detection via Sample Processing
title_full_unstemmed	Few-Shot Object Detection via Sample Processing
title_sort	few-shot object detection via sample processing
publisher	IEEE
publishDate	2021
url	https://doaj.org/article/73554ba2e8104e98bad6e1cc5b4a79f8
work_keys_str_mv	AT honghuixu fewshotobjectdetectionviasampleprocessing AT xinqingwang fewshotobjectdetectionviasampleprocessing AT famingshao fewshotobjectdetectionviasampleprocessing AT baoguoduan fewshotobjectdetectionviasampleprocessing AT pengzhang fewshotobjectdetectionviasampleprocessing
_version_	1718420674256568320

Few-Shot Object Detection via Sample Processing

Ejemplares similares