Scalable Object Detection for Edge Cloud Environments
Object detection is an important problem in a wide variety of computer vision applications for sustainable smart cities. Deep neural networks (DNNs) have attracted increasing interest in object detection due to their potential to provide high accuracy detection performance in challenging scenarios....
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2021-07-01
|
| Series: | Frontiers in Sustainable Cities |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/frsc.2021.675889/full |
| _version_ | 1818920896252870656 |
|---|---|
| author | Rory Hector Muhammad Umar Asif Mehmood Zhu Li Shuvra Bhattacharyya |
| author_facet | Rory Hector Muhammad Umar Asif Mehmood Zhu Li Shuvra Bhattacharyya |
| author_sort | Rory Hector |
| collection | DOAJ |
| description | Object detection is an important problem in a wide variety of computer vision applications for sustainable smart cities. Deep neural networks (DNNs) have attracted increasing interest in object detection due to their potential to provide high accuracy detection performance in challenging scenarios. However, DNNs involve high computational complexity and are therefore challenging to deploy under the tighter resource constraints of edge cloud environments compared to more resource-abundant platforms, such as conventional cloud computing platforms. Moreover, the monolithic structure of conventional DNN implementations limits their utility under the dynamically changing operational conditions that are typical in edge cloud computing. In this paper, we address these challenges and limitations of conventional DNN implementation techniques by introducing a new resource-adaptive scheme for DNN-based object detection. This scheme applies the recently-introduced concept of elastic neural networks, which involves the incorporation of multiple outputs within intermediate stages of the neural network backbone. We demonstrate a novel elastic DNN design for object detection, and we show how other methods for streamlining resource requirements, in particular network pruning, can be applied in conjunction with the proposed elastic network approach. Through extensive experiments, we demonstrate the ability of our methods to efficiently trade-off computational complexity and object detection accuracy for scalable deployment. |
| first_indexed | 2024-12-20T01:29:02Z |
| format | Article |
| id | doaj.art-0f47860ca1c74c5d9d5b5c27b6d5c250 |
| institution | Directory Open Access Journal |
| issn | 2624-9634 |
| language | English |
| last_indexed | 2024-12-20T01:29:02Z |
| publishDate | 2021-07-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Sustainable Cities |
| spelling | doaj.art-0f47860ca1c74c5d9d5b5c27b6d5c2502022-12-21T19:58:10ZengFrontiers Media S.A.Frontiers in Sustainable Cities2624-96342021-07-01310.3389/frsc.2021.675889675889Scalable Object Detection for Edge Cloud EnvironmentsRory Hector0Muhammad Umar1Asif Mehmood2Zhu Li3Shuvra Bhattacharyya4Division of Electrical and Computer Engineering, Louisiana State University, Baton Rouge, LA, United StatesDepartment of Electrical and Computer Engineering, University of Maryland, College Park, MD, United StatesU.S. Air Force Research Laboratory, Wright-Patterson Air Force Base, Dayton, OH, United StatesDepartment of Computer Science & Electrical Engnieering, University of Missouri, Kansas City, MO, United StatesDepartment of Electrical and Computer Engineering, University of Maryland, College Park, MD, United StatesObject detection is an important problem in a wide variety of computer vision applications for sustainable smart cities. Deep neural networks (DNNs) have attracted increasing interest in object detection due to their potential to provide high accuracy detection performance in challenging scenarios. However, DNNs involve high computational complexity and are therefore challenging to deploy under the tighter resource constraints of edge cloud environments compared to more resource-abundant platforms, such as conventional cloud computing platforms. Moreover, the monolithic structure of conventional DNN implementations limits their utility under the dynamically changing operational conditions that are typical in edge cloud computing. In this paper, we address these challenges and limitations of conventional DNN implementation techniques by introducing a new resource-adaptive scheme for DNN-based object detection. This scheme applies the recently-introduced concept of elastic neural networks, which involves the incorporation of multiple outputs within intermediate stages of the neural network backbone. We demonstrate a novel elastic DNN design for object detection, and we show how other methods for streamlining resource requirements, in particular network pruning, can be applied in conjunction with the proposed elastic network approach. Through extensive experiments, we demonstrate the ability of our methods to efficiently trade-off computational complexity and object detection accuracy for scalable deployment.https://www.frontiersin.org/articles/10.3389/frsc.2021.675889/fullobject detectiondeep learningelastic networksedge computingcomputer vision |
| spellingShingle | Rory Hector Muhammad Umar Asif Mehmood Zhu Li Shuvra Bhattacharyya Scalable Object Detection for Edge Cloud Environments Frontiers in Sustainable Cities object detection deep learning elastic networks edge computing computer vision |
| title | Scalable Object Detection for Edge Cloud Environments |
| title_full | Scalable Object Detection for Edge Cloud Environments |
| title_fullStr | Scalable Object Detection for Edge Cloud Environments |
| title_full_unstemmed | Scalable Object Detection for Edge Cloud Environments |
| title_short | Scalable Object Detection for Edge Cloud Environments |
| title_sort | scalable object detection for edge cloud environments |
| topic | object detection deep learning elastic networks edge computing computer vision |
| url | https://www.frontiersin.org/articles/10.3389/frsc.2021.675889/full |
| work_keys_str_mv | AT roryhector scalableobjectdetectionforedgecloudenvironments AT muhammadumar scalableobjectdetectionforedgecloudenvironments AT asifmehmood scalableobjectdetectionforedgecloudenvironments AT zhuli scalableobjectdetectionforedgecloudenvironments AT shuvrabhattacharyya scalableobjectdetectionforedgecloudenvironments |