TY - GEN
T1 - LIGHT-WEIGHT MIXED STAGE PARTIAL NETWORK FOR SURVEILLANCE OBJECT DETECTION WITH BACKGROUND DATA AUGMENTATION
AU - Ping-Yang, Chen
AU - Hsieh, Jun Wei
AU - Gochoo, Munkhjargal
AU - Chen, Yong Sheng
N1 - Funding Information:
This work was supported by a Grant-in-Aid for Scientific Research (14205046) from the Ministry of Education, Culture, Sports, Science and Technology. EF is indebted to Osaka University, Department of Electronic Engineering, for hospitality and acknowledges the support by Grant of Russian Fund for Basic Research 03-03-32507.
Publisher Copyright:
© 2021 IEEE.
PY - 2021
Y1 - 2021
N2 - State-of-the-art (SoTA) models have improved object detection accuracy with a large margin via convolutional neural networks, however still with an inferior performance for small objects. Moreover, these models are trained mainly based on the COCO dataset, and its backgrounds are more complicated than road environments, and thus degrade the accuracy of small road object detection. Compared with the COCO dataset, the background of a surveillance video is relatively stable and can be used to enhance the accuracy of road object detection. This paper designs a computationally efficient mixed stage partial (MSP) network to detect road objects. Another novelty of this paper is to propose a mixed background data augmentation method to enhance the detection accuracy without adding new labelling efforts. During inference, only the input image is used to detect road objects without further using any subtraction information. Extensive experiments on KITTI and UA-DETRAC benchmarks show the proposed method achieves the SoTA results for highly-accurate and efficient road object detection.
AB - State-of-the-art (SoTA) models have improved object detection accuracy with a large margin via convolutional neural networks, however still with an inferior performance for small objects. Moreover, these models are trained mainly based on the COCO dataset, and its backgrounds are more complicated than road environments, and thus degrade the accuracy of small road object detection. Compared with the COCO dataset, the background of a surveillance video is relatively stable and can be used to enhance the accuracy of road object detection. This paper designs a computationally efficient mixed stage partial (MSP) network to detect road objects. Another novelty of this paper is to propose a mixed background data augmentation method to enhance the detection accuracy without adding new labelling efforts. During inference, only the input image is used to detect road objects without further using any subtraction information. Extensive experiments on KITTI and UA-DETRAC benchmarks show the proposed method achieves the SoTA results for highly-accurate and efficient road object detection.
KW - Background subtraction
KW - Mixing Background Augmentation (MBA)
KW - MSPNet
KW - Road object detection
UR - http://www.scopus.com/inward/record.url?scp=85125575150&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85125575150&partnerID=8YFLogxK
U2 - 10.1109/ICIP42928.2021.9506212
DO - 10.1109/ICIP42928.2021.9506212
M3 - Conference contribution
AN - SCOPUS:85125575150
T3 - Proceedings - International Conference on Image Processing, ICIP
SP - 3333
EP - 3337
BT - 2021 IEEE International Conference on Image Processing, ICIP 2021 - Proceedings
PB - IEEE Computer Society
T2 - 2021 IEEE International Conference on Image Processing, ICIP 2021
Y2 - 19 September 2021 through 22 September 2021
ER -