LRP-DS: Lightweight RepPoints with Decoupled Sampling Point Set
Most object detection methods use rectangular bounding boxes to represent the object, while the representative points network (RepPoints) employs a point set to describe the object. The RepPoints can provide more fine-grained localization and facilitates classification. However, it ignores the diffe...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-06-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/11/13/5876 |
| _version_ | 1797528902960152576 |
|---|---|
| author | Jinchao Wang Libo Weng Fei Gao |
| author_facet | Jinchao Wang Libo Weng Fei Gao |
| author_sort | Jinchao Wang |
| collection | DOAJ |
| description | Most object detection methods use rectangular bounding boxes to represent the object, while the representative points network (RepPoints) employs a point set to describe the object. The RepPoints can provide more fine-grained localization and facilitates classification. However, it ignores the difference between localization and classification tasks. Therefore, a lightweight RepPoints with decoupling of the sampling point set (LRP-DS) is proposed in this paper. Firstly, the lightweight MobileNet-V2 and Feature Pyramid Networks (FPN) is employed as the backbone network to realize the lightweight network, rather than the Resnet. Secondly, considering the difference between classification and localization tasks, the sampling points of classification and localization are decoupled, by introducing classification free sampling method. Finally, due to the introduction of the classification free sampling method, the problem of the mismatch between the localization accuracy and the classification confidence is highlighted, so the localization score is employed to describe the localization accuracy independently. The final network structure of this paper achieves 73.3% mean average precision (mAP) on the VOC07 test dataset, which is 1.9% higher than original RepPoints with the same backbone network MobileNetV2 and FPN. Our LRP-DS has a detection speed of 20FPS for the input image of (1000, 600), on RTX2060 GPU, which is nearly twice as fast as the backbone network of ResNet50 and FPN. Experimental results show the effectiveness of our method. |
| first_indexed | 2024-03-10T10:05:56Z |
| format | Article |
| id | doaj.art-cda7663cba664c18bd958a15c12a9678 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-10T10:05:56Z |
| publishDate | 2021-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-cda7663cba664c18bd958a15c12a96782023-11-22T01:34:57ZengMDPI AGApplied Sciences2076-34172021-06-011113587610.3390/app11135876LRP-DS: Lightweight RepPoints with Decoupled Sampling Point SetJinchao Wang0Libo Weng1Fei Gao2Laboratory of Graphics & Image Processing, Zhejiang University of Technology, Xihu District, Hangzhou 310000, ChinaLaboratory of Graphics & Image Processing, Zhejiang University of Technology, Xihu District, Hangzhou 310000, ChinaLaboratory of Graphics & Image Processing, Zhejiang University of Technology, Xihu District, Hangzhou 310000, ChinaMost object detection methods use rectangular bounding boxes to represent the object, while the representative points network (RepPoints) employs a point set to describe the object. The RepPoints can provide more fine-grained localization and facilitates classification. However, it ignores the difference between localization and classification tasks. Therefore, a lightweight RepPoints with decoupling of the sampling point set (LRP-DS) is proposed in this paper. Firstly, the lightweight MobileNet-V2 and Feature Pyramid Networks (FPN) is employed as the backbone network to realize the lightweight network, rather than the Resnet. Secondly, considering the difference between classification and localization tasks, the sampling points of classification and localization are decoupled, by introducing classification free sampling method. Finally, due to the introduction of the classification free sampling method, the problem of the mismatch between the localization accuracy and the classification confidence is highlighted, so the localization score is employed to describe the localization accuracy independently. The final network structure of this paper achieves 73.3% mean average precision (mAP) on the VOC07 test dataset, which is 1.9% higher than original RepPoints with the same backbone network MobileNetV2 and FPN. Our LRP-DS has a detection speed of 20FPS for the input image of (1000, 600), on RTX2060 GPU, which is nearly twice as fast as the backbone network of ResNet50 and FPN. Experimental results show the effectiveness of our method.https://www.mdpi.com/2076-3417/11/13/5876RepPointsLRP-DSdecouplesamplemismatch |
| spellingShingle | Jinchao Wang Libo Weng Fei Gao LRP-DS: Lightweight RepPoints with Decoupled Sampling Point Set Applied Sciences RepPoints LRP-DS decouple sample mismatch |
| title | LRP-DS: Lightweight RepPoints with Decoupled Sampling Point Set |
| title_full | LRP-DS: Lightweight RepPoints with Decoupled Sampling Point Set |
| title_fullStr | LRP-DS: Lightweight RepPoints with Decoupled Sampling Point Set |
| title_full_unstemmed | LRP-DS: Lightweight RepPoints with Decoupled Sampling Point Set |
| title_short | LRP-DS: Lightweight RepPoints with Decoupled Sampling Point Set |
| title_sort | lrp ds lightweight reppoints with decoupled sampling point set |
| topic | RepPoints LRP-DS decouple sample mismatch |
| url | https://www.mdpi.com/2076-3417/11/13/5876 |
| work_keys_str_mv | AT jinchaowang lrpdslightweightreppointswithdecoupledsamplingpointset AT liboweng lrpdslightweightreppointswithdecoupledsamplingpointset AT feigao lrpdslightweightreppointswithdecoupledsamplingpointset |