An improved Tiny YOLOv3 for real-time object detection
The existing real-time object detection algorithm often omits the objects in the object detection. So an improved Tiny YOLOv3 (you look only once) algorithm is proposed with both lightweight and high accuracy of object detection. The improved Tiny YOLOv3 uses K-means clustering to estimate the size...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Taylor & Francis Group
2021-01-01
|
Series: | Systems Science & Control Engineering |
Subjects: | |
Online Access: | http://dx.doi.org/10.1080/21642583.2021.1901156 |
id |
doaj-2b3ab52a41634ba2930a5b7b4656a337 |
---|---|
record_format |
Article |
spelling |
doaj-2b3ab52a41634ba2930a5b7b4656a3372021-04-06T13:27:32ZengTaylor & Francis GroupSystems Science & Control Engineering2164-25832021-01-019131432110.1080/21642583.2021.19011561901156An improved Tiny YOLOv3 for real-time object detectionWendong Gai0Yakun Liu1Jing Zhang2Gang Jing3Shandong University of Science and TechnologyShandong University of Science and TechnologyShandong University of Science and TechnologyShandong University of Science and TechnologyThe existing real-time object detection algorithm often omits the objects in the object detection. So an improved Tiny YOLOv3 (you look only once) algorithm is proposed with both lightweight and high accuracy of object detection. The improved Tiny YOLOv3 uses K-means clustering to estimate the size of the anchor boxes for dataset. The pooling and convolution layers are added in the network to strengthen feature fusion and reduce parameters. The network structure increases upsampling and downsampling to enhance multi-scale fusion. The complete intersection over union is added in the loss function, which effectively improves the detection results. In addition, the proposed method has the lightweight module size and can be trained in the CPU. The experimental results show that the proposed method can meet the requirements of the detection speed and accuracy.http://dx.doi.org/10.1080/21642583.2021.1901156object detectiontiny yolov3multi-scale predictionk-meansreal-time |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Wendong Gai Yakun Liu Jing Zhang Gang Jing |
spellingShingle |
Wendong Gai Yakun Liu Jing Zhang Gang Jing An improved Tiny YOLOv3 for real-time object detection Systems Science & Control Engineering object detection tiny yolov3 multi-scale prediction k-means real-time |
author_facet |
Wendong Gai Yakun Liu Jing Zhang Gang Jing |
author_sort |
Wendong Gai |
title |
An improved Tiny YOLOv3 for real-time object detection |
title_short |
An improved Tiny YOLOv3 for real-time object detection |
title_full |
An improved Tiny YOLOv3 for real-time object detection |
title_fullStr |
An improved Tiny YOLOv3 for real-time object detection |
title_full_unstemmed |
An improved Tiny YOLOv3 for real-time object detection |
title_sort |
improved tiny yolov3 for real-time object detection |
publisher |
Taylor & Francis Group |
series |
Systems Science & Control Engineering |
issn |
2164-2583 |
publishDate |
2021-01-01 |
description |
The existing real-time object detection algorithm often omits the objects in the object detection. So an improved Tiny YOLOv3 (you look only once) algorithm is proposed with both lightweight and high accuracy of object detection. The improved Tiny YOLOv3 uses K-means clustering to estimate the size of the anchor boxes for dataset. The pooling and convolution layers are added in the network to strengthen feature fusion and reduce parameters. The network structure increases upsampling and downsampling to enhance multi-scale fusion. The complete intersection over union is added in the loss function, which effectively improves the detection results. In addition, the proposed method has the lightweight module size and can be trained in the CPU. The experimental results show that the proposed method can meet the requirements of the detection speed and accuracy. |
topic |
object detection tiny yolov3 multi-scale prediction k-means real-time |
url |
http://dx.doi.org/10.1080/21642583.2021.1901156 |
work_keys_str_mv |
AT wendonggai animprovedtinyyolov3forrealtimeobjectdetection AT yakunliu animprovedtinyyolov3forrealtimeobjectdetection AT jingzhang animprovedtinyyolov3forrealtimeobjectdetection AT gangjing animprovedtinyyolov3forrealtimeobjectdetection AT wendonggai improvedtinyyolov3forrealtimeobjectdetection AT yakunliu improvedtinyyolov3forrealtimeobjectdetection AT jingzhang improvedtinyyolov3forrealtimeobjectdetection AT gangjing improvedtinyyolov3forrealtimeobjectdetection |
_version_ |
1721538194537709568 |