We propose a target detection network (RMF-Net) based on the multi-scale strategy to solve the problems of large differences in the detection scale and mutual occlusion, which result in inaccurate locations. A multi-layer feature fusion module and multi-expansion dilated convolution pyramid module were designed based on the ResNet-101 residual network. The ability of the network to express the multi-scale features of the target could be improved by combining the shallow and deep features of the target and expanding the receptive field of the network. Moreover, RoI Align pooling was introduced to reduce the low accuracy of the anchor frame caused by multiple quantizations for improved positioning accuracy. Finally, an AD-IoU loss function was designed, which can adaptively optimise the distance between the prediction box and real box by comprehensively considering the overlap rate, centre distance, and aspect ratio between the boxes and can improve the detection accuracy of the occlusion target. Ablation experiments on the RMF-Net model verified the effectiveness of each factor in improving the network detection accuracy. Comparative experiments were conducted on the Pascal VOC2007 and Pascal VOC2012 datasets with various target detection algorithms based on convolutional neural networks. The results demonstrated that RMF-Net exhibited strong scale adaptability at different occlusion rates. The detection accuracy reached 80.4% and 78.5% respectively.
Yanyan ZHANG
Nanjing University of Information Science & Technology
Meiling SHEN
Nanjing University of Information Science & Technology
Wensheng YANG
Nanjing University of Information Science & Technology
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Yanyan ZHANG, Meiling SHEN, Wensheng YANG, "RMF-Net: Improving Object Detection with Multi-Scale Strategy" in IEICE TRANSACTIONS on Communications,
vol. E105-B, no. 5, pp. 675-683, May 2022, doi: 10.1587/transcom.2021EBP3102.
Abstract: We propose a target detection network (RMF-Net) based on the multi-scale strategy to solve the problems of large differences in the detection scale and mutual occlusion, which result in inaccurate locations. A multi-layer feature fusion module and multi-expansion dilated convolution pyramid module were designed based on the ResNet-101 residual network. The ability of the network to express the multi-scale features of the target could be improved by combining the shallow and deep features of the target and expanding the receptive field of the network. Moreover, RoI Align pooling was introduced to reduce the low accuracy of the anchor frame caused by multiple quantizations for improved positioning accuracy. Finally, an AD-IoU loss function was designed, which can adaptively optimise the distance between the prediction box and real box by comprehensively considering the overlap rate, centre distance, and aspect ratio between the boxes and can improve the detection accuracy of the occlusion target. Ablation experiments on the RMF-Net model verified the effectiveness of each factor in improving the network detection accuracy. Comparative experiments were conducted on the Pascal VOC2007 and Pascal VOC2012 datasets with various target detection algorithms based on convolutional neural networks. The results demonstrated that RMF-Net exhibited strong scale adaptability at different occlusion rates. The detection accuracy reached 80.4% and 78.5% respectively.
URL: https://global.ieice.org/en_transactions/communications/10.1587/transcom.2021EBP3102/_p
Copy
@ARTICLE{e105-b_5_675,
author={Yanyan ZHANG, Meiling SHEN, Wensheng YANG, },
journal={IEICE TRANSACTIONS on Communications},
title={RMF-Net: Improving Object Detection with Multi-Scale Strategy},
year={2022},
volume={E105-B},
number={5},
pages={675-683},
abstract={We propose a target detection network (RMF-Net) based on the multi-scale strategy to solve the problems of large differences in the detection scale and mutual occlusion, which result in inaccurate locations. A multi-layer feature fusion module and multi-expansion dilated convolution pyramid module were designed based on the ResNet-101 residual network. The ability of the network to express the multi-scale features of the target could be improved by combining the shallow and deep features of the target and expanding the receptive field of the network. Moreover, RoI Align pooling was introduced to reduce the low accuracy of the anchor frame caused by multiple quantizations for improved positioning accuracy. Finally, an AD-IoU loss function was designed, which can adaptively optimise the distance between the prediction box and real box by comprehensively considering the overlap rate, centre distance, and aspect ratio between the boxes and can improve the detection accuracy of the occlusion target. Ablation experiments on the RMF-Net model verified the effectiveness of each factor in improving the network detection accuracy. Comparative experiments were conducted on the Pascal VOC2007 and Pascal VOC2012 datasets with various target detection algorithms based on convolutional neural networks. The results demonstrated that RMF-Net exhibited strong scale adaptability at different occlusion rates. The detection accuracy reached 80.4% and 78.5% respectively.},
keywords={},
doi={10.1587/transcom.2021EBP3102},
ISSN={1745-1345},
month={May},}
Copy
TY - JOUR
TI - RMF-Net: Improving Object Detection with Multi-Scale Strategy
T2 - IEICE TRANSACTIONS on Communications
SP - 675
EP - 683
AU - Yanyan ZHANG
AU - Meiling SHEN
AU - Wensheng YANG
PY - 2022
DO - 10.1587/transcom.2021EBP3102
JO - IEICE TRANSACTIONS on Communications
SN - 1745-1345
VL - E105-B
IS - 5
JA - IEICE TRANSACTIONS on Communications
Y1 - May 2022
AB - We propose a target detection network (RMF-Net) based on the multi-scale strategy to solve the problems of large differences in the detection scale and mutual occlusion, which result in inaccurate locations. A multi-layer feature fusion module and multi-expansion dilated convolution pyramid module were designed based on the ResNet-101 residual network. The ability of the network to express the multi-scale features of the target could be improved by combining the shallow and deep features of the target and expanding the receptive field of the network. Moreover, RoI Align pooling was introduced to reduce the low accuracy of the anchor frame caused by multiple quantizations for improved positioning accuracy. Finally, an AD-IoU loss function was designed, which can adaptively optimise the distance between the prediction box and real box by comprehensively considering the overlap rate, centre distance, and aspect ratio between the boxes and can improve the detection accuracy of the occlusion target. Ablation experiments on the RMF-Net model verified the effectiveness of each factor in improving the network detection accuracy. Comparative experiments were conducted on the Pascal VOC2007 and Pascal VOC2012 datasets with various target detection algorithms based on convolutional neural networks. The results demonstrated that RMF-Net exhibited strong scale adaptability at different occlusion rates. The detection accuracy reached 80.4% and 78.5% respectively.
ER -