{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,10,12]],"date-time":"2025-10-12T02:04:51Z","timestamp":1760234691811,"version":"build-2065373602"},"reference-count":34,"publisher":"MDPI AG","issue":"12","license":[{"start":{"date-parts":[[2021,6,14]],"date-time":"2021-06-14T00:00:00Z","timestamp":1623628800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100014188","name":"Ministry of Science and ICT, South Korea","doi-asserted-by":"publisher","award":["2018-0-01502"],"award-info":[{"award-number":["2018-0-01502"]}],"id":[{"id":"10.13039\/501100014188","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Gachon University research fund of 2020","award":["GCU-202008450004"],"award-info":[{"award-number":["GCU-202008450004"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Sensors"],"abstract":"<jats:p>This paper presents a novel adaptive object movement and motion tracking (AdaMM) framework in a hierarchical edge computing system for achieving GPU memory footprint reduction of deep learning (DL)-based video surveillance services. DL-based object movement and motion tracking requires a significant amount of resources, such as (1) GPU processing power for the inference phase and (2) GPU memory for model loading. Despite the absence of an object in the video, if the DL model is loaded, the GPU memory must be kept allocated for the loaded model. Moreover, in several cases, video surveillance tries to capture events that rarely occur (e.g., abnormal object behaviors); therefore, such standby GPU memory might be easily wasted. To alleviate this problem, the proposed AdaMM framework categorizes the tasks used for the object movement and motion tracking procedure in an increasing order of the required processing and memory resources as task (1) frame difference calculation, task (2) object detection, and task (3) object motion and movement tracking. The proposed framework aims to adaptively release the unnecessary standby object motion and movement tracking model to save GPU memory by utilizing light tasks, such as frame difference calculation and object detection in a hierarchical manner. Consequently, object movement and motion tracking are adaptively triggered if the object is detected within the specified threshold time; otherwise, the GPU memory for the model of task (3) can be released. Moreover, object detection is also adaptively performed if the frame difference over time is greater than the specified threshold. We implemented the proposed AdaMM framework using commercial edge devices by considering a three-tier system, such as the 1st edge node for both tasks (1) and (2), the 2nd edge node for task (3), and the cloud for sending a push alarm. A measurement-based experiment reveals that the proposed framework achieves a maximum GPU memory reduction of 76.8% compared to the baseline system, while requiring a 2680 ms delay for loading the model for object movement and motion tracking.<\/jats:p>","DOI":"10.3390\/s21124089","type":"journal-article","created":{"date-parts":[[2021,6,14]],"date-time":"2021-06-14T22:25:46Z","timestamp":1623709546000},"page":"4089","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["AdaMM: Adaptive Object Movement and Motion Tracking in Hierarchical Edge Computing System"],"prefix":"10.3390","volume":"21","author":[{"given":"Jingyeom","family":"Kim","sequence":"first","affiliation":[{"name":"School of Computing, Gachon University, Seongnam 13120, Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1102-3905","authenticated-orcid":false,"given":"Joohyung","family":"Lee","sequence":"additional","affiliation":[{"name":"School of Computing, Gachon University, Seongnam 13120, Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Taeyeon","family":"Kim","sequence":"additional","affiliation":[{"name":"Electronics and Telecommunications Research Institute (ETRI), Daejeon 34129, Korea"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1968","published-online":{"date-parts":[[2021,6,14]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"609","DOI":"10.32604\/iasc.2020.013939","article-title":"Research on the Automatic Extraction Method of Web Data Objects Based on Deep Learning","volume":"26","author":"Peng","year":"2020","journal-title":"Intell. Autom. Soft Comput."},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"887","DOI":"10.32604\/iasc.2020.010122","article-title":"Multi-scale boxes loss for object detection in smart energy","volume":"26","author":"Dai","year":"2020","journal-title":"Intell. Autom. Soft Comput."},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"465","DOI":"10.1016\/j.procs.2018.08.198","article-title":"Design of smart home security system using object recognition and PIR sensor","volume":"135","author":"Surantha","year":"2018","journal-title":"Procedia Comput. Sci."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"2031","DOI":"10.1109\/COMST.2020.2986024","article-title":"Federated learning in mobile edge networks: A comprehensive survey","volume":"22","author":"Lim","year":"2020","journal-title":"IEEE Commun. Surv. Tutor."},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"1738","DOI":"10.1109\/JPROC.2019.2918951","article-title":"Edge intelligence: Paving the last mile of artificial intelligence with edge computing","volume":"107","author":"Zhou","year":"2019","journal-title":"Proc. IEEE"},{"key":"ref_6","first-page":"569","article-title":"User profile system based on sentiment analysis for mobile edge computing","volume":"62","author":"Kim","year":"2020","journal-title":"Comput. Mater. Contin."},{"key":"ref_7","first-page":"905","article-title":"Edge computing-based tasks offloading and block caching for mobile blockchain","volume":"62","author":"Dai","year":"2020","journal-title":"Comput. Mater. Contin."},{"key":"ref_8","first-page":"1033","article-title":"DL-HAR: Deep Learning-Based Human Activity Recognition Framework for Edge Computing","volume":"65","author":"Gumaei","year":"2020","journal-title":"CMC-Comput. Mater. Contin."},{"key":"ref_9","unstructured":"(2021, April 05). Device Development for 5G Era and beyond. Available online: https:\/\/www.nttdocomo.co.jp\/info\/news_release\/2019\/08\/29_01.html."},{"key":"ref_10","unstructured":"(2021, April 05). Video \u00d7 5G to Accelerate Digital Transformation to Enterprise 5G-Enabled. Available online: https:\/\/news.kddi.com\/kddi\/corporate\/newsrelease\/2019\/09\/02\/3995.html."},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Chen, T.Y.H., Ravindranath, L., Deng, S., Bahl, P., and Balakrishnan, H. (2015, January 1\u20134). Glimpse: Continuous, real-time object recognition on mobile devices. Proceedings of the 13th ACM Conference on Embedded Networked Sensor Systems, Seoul, Korea.","DOI":"10.1145\/2809695.2809711"},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Drolia, U., Guo, K., Tan, J., Gandhi, R., and Narasimhan, P. (2017, January 5\u20138). Cachier: Edge-caching for recognition applications. Proceedings of the 2017 IEEE 37th International Conference on Distributed Computing Systems (ICDCS), Atlanta, GA, USA.","DOI":"10.1109\/ICDCS.2017.94"},{"key":"ref_13","doi-asserted-by":"crossref","unstructured":"Teerapittayanon, S., McDanel, B., and Kung, H.T. (2016, January 4\u20138). Branchynet: Fast inference via early exiting from deep neural networks. Proceedings of the 2016 23rd International Conference on Pattern Recognition (ICPR), Cancun, Mexico.","DOI":"10.1109\/ICPR.2016.7900006"},{"key":"ref_14","doi-asserted-by":"crossref","unstructured":"Li, E., Zhou, Z., and Chen, X. (2018, January 20). Edge intelligence: On-demand deep learning model co-inference with device-edge synergy. Proceedings of the 2018 Workshop on Mobile Edge Communications, Budapest, Hungary.","DOI":"10.1145\/3229556.3229562"},{"key":"ref_15","unstructured":"Han, S., Mao, H., and Dally, W.J. (2015). Deep compression: Compressing deep neural networks with pruning, trained quantization and huffman coding. arXiv."},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"179104","DOI":"10.1109\/ACCESS.2019.2958406","article-title":"A Robust Deep-Neural-Network-Based Compressed Model for Mobile Device Assisted by Edge Server","volume":"7","author":"Yan","year":"2019","journal-title":"IEEE Access"},{"key":"ref_17","doi-asserted-by":"crossref","unstructured":"Li, H., Hu, C., Jiang, J., Wang, Z., Wen, Y., and Zhu, W. (2018, January 11\u201313). Jalad: Joint accuracy-and latency-aware deep structure decoupling for edge-cloud execution. Proceedings of the 2018 IEEE 24th International Conference on Parallel and Distributed Systems (ICPADS), Singapore.","DOI":"10.1109\/PADSW.2018.8645013"},{"key":"ref_18","doi-asserted-by":"crossref","unstructured":"Hu, C., Bao, W., Wang, D., and Liu, F. (May, January 29). Dynamic adaptive DNN surgery for inference acceleration on the edge. Proceedings of the IEEE INFOCOM 2019-IEEE Conference on Computer Communications, Paris, France.","DOI":"10.1109\/INFOCOM.2019.8737614"},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"35119","DOI":"10.1007\/s11042-019-08067-1","article-title":"UAV based cost-effective real-time abnormal event detection using edge computing","volume":"78","author":"Alam","year":"2019","journal-title":"Multimed. Tools Appl."},{"key":"ref_20","first-page":"205","article-title":"Visual object detection and tracking using analytical learning approach of validity level","volume":"25","author":"Lee","year":"2019","journal-title":"Intell. Autom. Soft Comput."},{"key":"ref_21","doi-asserted-by":"crossref","unstructured":"Lee, S., and Nirjon, S. (2020, January 15\u201319). Fast and scalable in-memory deep multitask learning via neural weight virtualization. Proceedings of the 18th International Conference on Mobile Systems, Applications, and Services, Toronto, ON, Canada.","DOI":"10.1145\/3386901.3388947"},{"key":"ref_22","unstructured":"Meng, C., Sun, M., Yang, J., Qiu, M., and Gu, Y. (2017, January 8). Training deeper models by GPU memory optimization on TensorFlow. Proceedings of the ML Systems Workshop in NIPS, Long Beach, CA, USA."},{"key":"ref_23","doi-asserted-by":"crossref","first-page":"2740","DOI":"10.1109\/TSMC.2017.2690673","article-title":"Gpu-accelerated parallel hierarchical extreme learning machine on flink for big data","volume":"47","author":"Chen","year":"2017","journal-title":"IEEE Trans. Syst. Man Cybern. Syst."},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"2623","DOI":"10.1109\/TC.2014.2366731","article-title":"Performance optimization using partitioned SpMV on GPUs and multicore CPUs","volume":"64","author":"Yang","year":"2014","journal-title":"IEEE Trans. Comput."},{"key":"ref_25","unstructured":"(2021, April 05). NVIDIA, CUDA. Available online: https:\/\/developer.nvidia.com\/cuda-zone."},{"key":"ref_26","unstructured":"(2020, August 20). OpenCV Color Conversions. Available online: https:\/\/docs.opencv.org\/master\/de\/d25\/imgproc_color_conversions.html."},{"key":"ref_27","unstructured":"Redmon, J., and Farhadi, A. (2018). Yolov3: An incremental improvement. arXiv."},{"key":"ref_28","doi-asserted-by":"crossref","unstructured":"Tajar, A.T., Ramazani, A., and Mansoorizadeh, M. (2021). A lightweight Tiny-YOLOv3 vehicle detection approach. J. Real-Time Image Process., 1\u201313.","DOI":"10.1007\/s11554-021-01131-w"},{"key":"ref_29","unstructured":"(2021, June 04). NVIDIA Jetson Nano Devloper Kit. Available online: https:\/\/developer.nvidia.com\/embedded\/jetson-nano-developer-kit."},{"key":"ref_30","unstructured":"(2021, June 04). opencv-python4.2.0. Available online: https:\/\/pypi.org\/project\/opencv-python\/4.2.0.34\/."},{"key":"ref_31","unstructured":"(2020, August 20). tf-pose-estimation. Available online: https:\/\/github.com\/tryagainconcepts\/tf-pose-estimation."},{"key":"ref_32","unstructured":"(2020, August 20). Flask. Available online: https:\/\/flask.palletsprojects.com\/en\/1.1.x\/."},{"key":"ref_33","unstructured":"(2020, August 20). Android WebView. Available online: https:\/\/developer.android.com\/reference\/android\/webkit\/WebView."},{"key":"ref_34","unstructured":"(2020, August 20). Firebase Cloud Messaging. Available online: https:\/\/firebase.google.com\/docs\/cloud-messaging."}],"container-title":["Sensors"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1424-8220\/21\/12\/4089\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,11]],"date-time":"2025-10-11T06:14:02Z","timestamp":1760163242000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1424-8220\/21\/12\/4089"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,6,14]]},"references-count":34,"journal-issue":{"issue":"12","published-online":{"date-parts":[[2021,6]]}},"alternative-id":["s21124089"],"URL":"https:\/\/doi.org\/10.3390\/s21124089","relation":{},"ISSN":["1424-8220"],"issn-type":[{"type":"electronic","value":"1424-8220"}],"subject":[],"published":{"date-parts":[[2021,6,14]]}}}