High-resolution processing and sigmoid fusion modules for efficient detection of small objects in an embedded system
Abstract Recent advances in deep learning realized accurate, robust detection of various types of objects including pedestrians on the road, defect regions in the manufacturing process, human organs in medical images, and dangerous materials passing through the airport checkpoint. Specifically, smal...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2023-01-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-022-27189-5 |
| _version_ | 1797958695381893120 |
|---|---|
| author | Mingi Kim Heegwang Kim Junghoon Sung Chanyeong Park Joonki Paik |
| author_facet | Mingi Kim Heegwang Kim Junghoon Sung Chanyeong Park Joonki Paik |
| author_sort | Mingi Kim |
| collection | DOAJ |
| description | Abstract Recent advances in deep learning realized accurate, robust detection of various types of objects including pedestrians on the road, defect regions in the manufacturing process, human organs in medical images, and dangerous materials passing through the airport checkpoint. Specifically, small object detection implemented as an embedded system is gaining increasing attention for autonomous vehicles, drone reconnaissance, and microscopic imagery. In this paper, we present a light-weight small object detection model using two plug-in modules: (1) high-resolution processing module (HRPM ) and (2) sigmoid fusion module (SFM). The HRPM efficiently learns multi-scale features of small objects using a significantly reduced computational cost, and the SFM alleviates mis-classification errors due to spatial noise by adjusting weights on the lost small object information. Combination of HRPM and SFM significantly improved the detection accuracy with a low amount of computation. Compared with the original YOLOX-s model, the proposed model takes a two-times higher-resolution input image for higher mean average precision (mAP) using 57% model parameters and 71% computation in Gflops. The proposed model was tested using real drone reconnaissance images, and provided significant improvement in detecting small vehicles. |
| first_indexed | 2024-04-11T00:22:38Z |
| format | Article |
| id | doaj.art-25c30f2df71a431dbc397b1cc3e297e6 |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-04-11T00:22:38Z |
| publishDate | 2023-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-25c30f2df71a431dbc397b1cc3e297e62023-01-08T12:09:55ZengNature PortfolioScientific Reports2045-23222023-01-0113111210.1038/s41598-022-27189-5High-resolution processing and sigmoid fusion modules for efficient detection of small objects in an embedded systemMingi Kim0Heegwang Kim1Junghoon Sung2Chanyeong Park3Joonki Paik4Department of Artificial Intelligence, Chung-Ang UniversityDepartment of Image, Chung-Ang UniversityDepartment of Image, Chung-Ang UniversityDepartment of Image, Chung-Ang UniversityDepartment of Artificial Intelligence, Chung-Ang UniversityAbstract Recent advances in deep learning realized accurate, robust detection of various types of objects including pedestrians on the road, defect regions in the manufacturing process, human organs in medical images, and dangerous materials passing through the airport checkpoint. Specifically, small object detection implemented as an embedded system is gaining increasing attention for autonomous vehicles, drone reconnaissance, and microscopic imagery. In this paper, we present a light-weight small object detection model using two plug-in modules: (1) high-resolution processing module (HRPM ) and (2) sigmoid fusion module (SFM). The HRPM efficiently learns multi-scale features of small objects using a significantly reduced computational cost, and the SFM alleviates mis-classification errors due to spatial noise by adjusting weights on the lost small object information. Combination of HRPM and SFM significantly improved the detection accuracy with a low amount of computation. Compared with the original YOLOX-s model, the proposed model takes a two-times higher-resolution input image for higher mean average precision (mAP) using 57% model parameters and 71% computation in Gflops. The proposed model was tested using real drone reconnaissance images, and provided significant improvement in detecting small vehicles.https://doi.org/10.1038/s41598-022-27189-5 |
| spellingShingle | Mingi Kim Heegwang Kim Junghoon Sung Chanyeong Park Joonki Paik High-resolution processing and sigmoid fusion modules for efficient detection of small objects in an embedded system Scientific Reports |
| title | High-resolution processing and sigmoid fusion modules for efficient detection of small objects in an embedded system |
| title_full | High-resolution processing and sigmoid fusion modules for efficient detection of small objects in an embedded system |
| title_fullStr | High-resolution processing and sigmoid fusion modules for efficient detection of small objects in an embedded system |
| title_full_unstemmed | High-resolution processing and sigmoid fusion modules for efficient detection of small objects in an embedded system |
| title_short | High-resolution processing and sigmoid fusion modules for efficient detection of small objects in an embedded system |
| title_sort | high resolution processing and sigmoid fusion modules for efficient detection of small objects in an embedded system |
| url | https://doi.org/10.1038/s41598-022-27189-5 |
| work_keys_str_mv | AT mingikim highresolutionprocessingandsigmoidfusionmodulesforefficientdetectionofsmallobjectsinanembeddedsystem AT heegwangkim highresolutionprocessingandsigmoidfusionmodulesforefficientdetectionofsmallobjectsinanembeddedsystem AT junghoonsung highresolutionprocessingandsigmoidfusionmodulesforefficientdetectionofsmallobjectsinanembeddedsystem AT chanyeongpark highresolutionprocessingandsigmoidfusionmodulesforefficientdetectionofsmallobjectsinanembeddedsystem AT joonkipaik highresolutionprocessingandsigmoidfusionmodulesforefficientdetectionofsmallobjectsinanembeddedsystem |