{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,9,28]],"date-time":"2025-09-28T06:24:08Z","timestamp":1759040648889,"version":"3.41.2"},"reference-count":61,"publisher":"Association for Computing Machinery (ACM)","issue":"3","funder":[{"DOI":"10.13039\/501100001321","name":"National Research Foundation","doi-asserted-by":"crossref","award":["NRFNRFI05-2019-0007"],"award-info":[{"award-number":["NRFNRFI05-2019-0007"]}],"id":[{"id":"10.13039\/501100001321","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Trans. Internet Things"],"published-print":{"date-parts":[[2025,8,31]]},"abstract":"<jats:p>\n            While DNN models have dramatically improved the accuracy of machine vision tasks, pervasive deployments of vision sensors for surveillance tasks continue to suffer from high energy consumption and network traffic overhead. To tackle these problems, we introduce\n            <jats:italic toggle=\"yes\">CollabCam<\/jats:italic>\n            , an edge-based machine vision system designed for multi-camera deployments that leverages the naturally-occurring overlaps in the field-of-view (FoV) among neighboring cameras.\n            <jats:italic toggle=\"yes\">CollabCam<\/jats:italic>\n            synergistically combines two innovative ideas: (a) having each individual camera compose and transmit mixed-resolution frames (MRF) via lightweight down-sampling, where the transmitted images have significantly lower resolution in the\n            <jats:italic toggle=\"yes\">shared, overlapping portions<\/jats:italic>\n            , and (b) performing inference, for an exemplar object detection task, for each camera stream using a new collaborative mechanism which utilizes suitably-translated object bounding boxes from a peer \u201ccollaborating\u201d camera as an additional input channel. We demonstrate how this collaborative mechanism is generalizable and can be realized by simply retraining off-the-shelf object detector DNNs, such as YOLOv3 and SSD, without modifying their model structures. By emulating the performance of\n            <jats:italic toggle=\"yes\">CollabCam<\/jats:italic>\n            using two benchmark outdoor-campus multi-camera datasets, we show that\n            <jats:italic toggle=\"yes\">Collab-DNNs<\/jats:italic>\n            can accommodate a 50\u201360 fold reduction in image size (therefore reducing network transmission overhead), for both high-resolution (1056x1056) and low-resolution (512x512) images, with a modest \u2264 2 - 5% drop in object detection accuracy, compared to a non-collaborative approach that suffers a \u223c 45\u201360% drop in accuracy. Subsequently, by deploying a Raspberry-Pi based\n            <jats:italic toggle=\"yes\">CollabCam<\/jats:italic>\n            prototype on a campus-based test-bed, we demonstrate that\n            <jats:italic toggle=\"yes\">CollabCam<\/jats:italic>\n            can reduce the overall energy\/image frame overhead by \u223c25\u201335%, with even higher energy savings (\u223c35\u201345%) likely with hardware optimization. Finally, additional experiments help demonstrate that\n            <jats:italic toggle=\"yes\">CollabCam<\/jats:italic>\n            can prove beneficial for varied (including multi-class) object detection tasks and that\n            <jats:italic toggle=\"yes\">CollabCam\u2019s<\/jats:italic>\n            performance benefits may be best realized by ensuring that the number of deployed, collaborating cameras is not excessively high.\n          <\/jats:p>","DOI":"10.1145\/3736420","type":"journal-article","created":{"date-parts":[[2025,5,17]],"date-time":"2025-05-17T07:33:41Z","timestamp":1747467221000},"page":"1-35","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["CollabCam: Collaborative Inference and Mixed-Resolution Imaging for Energy-Efficient Pervasive Vision"],"prefix":"10.1145","volume":"6","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-0030-5210","authenticated-orcid":false,"given":"Amit","family":"Sharma","sequence":"first","affiliation":[{"name":"Singapore Management University School of Computing and Information Systems","place":["Singapore, Singapore"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-4815-7784","authenticated-orcid":false,"given":"Vithurson","family":"Subasharan","sequence":"additional","affiliation":[{"name":"Singapore Management University","place":["Singapore, Singapore"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0733-2510","authenticated-orcid":false,"given":"Manoj","family":"Gulati","sequence":"additional","affiliation":[{"name":"Singapore Management University","place":["Singapore, Singapore"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-5270-1794","authenticated-orcid":false,"given":"Dhanuja","family":"Wanniarachchi","sequence":"additional","affiliation":[{"name":"Singapore Management University","place":["Singapore, Singapore"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1212-1769","authenticated-orcid":false,"given":"Archan","family":"Misra","sequence":"additional","affiliation":[{"name":"School of Computing and Information Systems, Singapore Management University","place":["Singapore, Singapore"]}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"320","published-online":{"date-parts":[[2025,7,10]]},"reference":[{"doi-asserted-by":"publisher","unstructured":"Tarek Abdelzaher Yifan Hao Kasthuri Jayarajah Archan Misra Per Skarin Shuochao Yao Dulanga Weerakoon and Karl-Erik \u00c5rz\u00e9n. 2020. Five challenges in cloud-enabled intelligence and control. ACM Trans. Internet Technol. 20 1 (February 2020). DOI:10.1145\/3366021","key":"e_1_3_3_2_2","DOI":"10.1145\/3366021"},{"doi-asserted-by":"publisher","key":"e_1_3_3_3_2","DOI":"10.1109\/ICASSP.2007.367335"},{"unstructured":"Christopher Canel Thomas Kim Giulio Zhou Conglong Li Hyeontaek Lim David G. Andersen Michael Kaminsky and Subramanya Dulloor. 2019. Scaling video analytics on constrained edge nodes. Proceedings of Machine Learning and Systems 1 (2019) 406\u2013417.","key":"e_1_3_3_4_2"},{"doi-asserted-by":"publisher","key":"e_1_3_3_5_2","DOI":"10.1109\/CVPR.2018.00528"},{"doi-asserted-by":"publisher","unstructured":"Tatjana Chavdarova and Francois Fleuret. 2017. Deep multi-camera people detection. In 2017 16th IEEE International Conference on Machine Learning and Applications (ICMLA) 848\u2013853. DOI:10.1109\/ICMLA.2017.00-50","key":"e_1_3_3_6_2","DOI":"10.1109\/ICMLA.2017.00-50"},{"doi-asserted-by":"publisher","key":"e_1_3_3_7_2","DOI":"10.1109\/ICDCS.2017.152"},{"unstructured":"Salehe Erfanian Ebadi Saurav Dhakad Sanjay Vishwakarma Chunpu Wang You-Cyuan Jhang Maciek Chociej Adam Crespi Alex Thaman and Sujoy Ganguly. 2022. PSP-HDRI+: A synthetic dataset generator for pre-training of human-centric computer vision models. In First Workshop on Pre-training: Perspectives Pitfalls and Paths Forward at ICML 2022.","key":"e_1_3_3_8_2"},{"key":"e_1_3_3_9_2","article-title":"ESP32-CAM Development Board","year":"2022","unstructured":"Expressif_Systems. 2022. ESP32-CAM Development Board. Retrieved from https:\/\/media.digikey.com\/pdf\/Data%20Sheets\/DFRobot%20PDFs\/DFR0602_Web.pdf. (2022).","journal-title":"https:\/\/media.digikey.com\/pdf\/Data%20Sheets\/DFRobot%20PDFs\/DFR0602_Web.pdf"},{"doi-asserted-by":"publisher","key":"e_1_3_3_10_2","DOI":"10.1109\/PETS-WINTER.2009.5399556"},{"doi-asserted-by":"publisher","key":"e_1_3_3_11_2","DOI":"10.1109\/CVPR.2018.00724"},{"doi-asserted-by":"publisher","key":"e_1_3_3_12_2","DOI":"10.1007\/978-3-642-15555-0_24"},{"doi-asserted-by":"publisher","key":"e_1_3_3_13_2","DOI":"10.1145\/3587819.3590986"},{"doi-asserted-by":"publisher","key":"e_1_3_3_14_2","DOI":"10.1145\/3458305.3463381"},{"doi-asserted-by":"publisher","key":"e_1_3_3_15_2","DOI":"10.1109\/WACV56688.2023.00484"},{"unstructured":"Alexander Hermans Lucas Beyer and Bastian Leibe. 2017. In defense of the triplet loss for person re-identification. Retrieved from https:\/\/arxiv.org\/abs\/1703.07737","key":"e_1_3_3_16_2"},{"key":"e_1_3_3_17_2","article-title":"Homography Matrix for Coordinate Mapping","year":"2022","unstructured":"Homography_Matrix. 2022. Homography Matrix for Coordinate Mapping. Retrieved from https:\/\/docs.opencv.org\/master\/d9\/dab\/tutorial_homography.html. (2022).","journal-title":"https:\/\/docs.opencv.org\/master\/d9\/dab\/tutorial_homography.html"},{"doi-asserted-by":"publisher","key":"e_1_3_3_18_2","DOI":"10.1145\/3307334.3328594"},{"doi-asserted-by":"publisher","key":"e_1_3_3_19_2","DOI":"10.1145\/3301293.3302366"},{"doi-asserted-by":"publisher","key":"e_1_3_3_20_2","DOI":"10.1109\/SEC50012.2020.00016"},{"doi-asserted-by":"publisher","key":"e_1_3_3_21_2","DOI":"10.1145\/3230543.3230574"},{"doi-asserted-by":"publisher","unstructured":"Glenn Jocher Ayush Chaurasia Alex Stoken Jirka Borovec NanoCode012 tkianai yxNONG Piotr Skalski Adam Hogan Dhruv Nair Max Strobel and Mrinal Jain. 2022. ultralytics\/yolov5: v7.0 - YOLOv5 SOTA Realtime Instance Segmentation. (Nov.2022). DOI:DOI:10.5281\/zenodo.7347926","key":"e_1_3_3_22_2","DOI":"10.5281\/zenodo.7347926"},{"doi-asserted-by":"publisher","unstructured":"Daniel Kang John Emmons Firas Abuzaid Peter Bailis and Matei Zaharia. 2017. NoScope: optimizing neural network queries over video at scale. Proc. VLDB Endow. 10 11 (August 2017) 1586\u20131597. DOI:10.14778\/3137628.3137664","key":"e_1_3_3_23_2","DOI":"10.14778\/3137628.3137664"},{"doi-asserted-by":"publisher","key":"e_1_3_3_24_2","DOI":"10.1145\/3445814.3446737"},{"doi-asserted-by":"publisher","key":"e_1_3_3_25_2","DOI":"10.1145\/1101149.1101191"},{"doi-asserted-by":"publisher","key":"e_1_3_3_26_2","DOI":"10.1109\/AFGR.2002.1004148"},{"doi-asserted-by":"publisher","key":"e_1_3_3_27_2","DOI":"10.1145\/2462456.2464448"},{"doi-asserted-by":"publisher","key":"e_1_3_3_28_2","DOI":"10.1007\/978-3-319-46448-0_2"},{"doi-asserted-by":"publisher","key":"e_1_3_3_29_2","DOI":"10.1145\/3356250.3360041"},{"key":"e_1_3_3_30_2","article-title":"Monsoon Power Monitor Manual","year":"2022","unstructured":"Monsoon_Solutions. 2022. Monsoon Power Monitor Manual. Retrieved from http:\/\/msoon.github.io\/powermonitor\/PowerTool\/doc\/LVPM%20Manual.pdf. (2022).","journal-title":"http:\/\/msoon.github.io\/powermonitor\/PowerTool\/doc\/LVPM%20Manual.pdf"},{"doi-asserted-by":"publisher","key":"e_1_3_3_31_2","DOI":"10.1145\/3081333.3081347"},{"doi-asserted-by":"publisher","key":"e_1_3_3_32_2","DOI":"10.1109\/CVPRW53098.2021.00482"},{"key":"e_1_3_3_33_2","article-title":"NI USB-6003 Multifunction I\/O Device","year":"2022","unstructured":"National_Instruments. 2022. NI USB-6003 Multifunction I\/O Device. Retrieved from https:\/\/www.ni.com\/en-us\/support\/model.usb-6003.html. (2022).","journal-title":"https:\/\/www.ni.com\/en-us\/support\/model.usb-6003.html"},{"key":"e_1_3_3_34_2","article-title":"ARX3A0: CMOS Image Sensor, Ultra-Low-Power, 560 x 560, 360 fps","year":"2022","unstructured":"OnSemi_Image_Sensor. 2022. ARX3A0: CMOS Image Sensor, Ultra-Low-Power, 560 x 560, 360 fps. Retrieved from https:\/\/www.onsemi.com\/products\/sensors\/image-sensors\/arx3a0. (2022).","journal-title":"https:\/\/www.onsemi.com\/products\/sensors\/image-sensors\/arx3a0"},{"doi-asserted-by":"publisher","key":"e_1_3_3_35_2","DOI":"10.1016\/j.patcog.2014.12.004"},{"key":"e_1_3_3_36_2","article-title":"Official PiCamera Library","year":"2016","unstructured":"PiCAM_Lib. 2016. Official PiCamera Library. Retrieved from https:\/\/picamera.readthedocs.io\/en\/release-1.13\/. (2016).","journal-title":"https:\/\/picamera.readthedocs.io\/en\/release-1.13\/"},{"key":"e_1_3_3_37_2","article-title":"Raspberry Pi Camera Module v2","year":"2022","unstructured":"PiCAMv2. 2022. Raspberry Pi Camera Module v2. Retrieved from https:\/\/www.raspberrypi.org\/products\/camera-module-v2\/?resellerType=home. (2022).","journal-title":"https:\/\/www.raspberrypi.org\/products\/camera-module-v2\/?resellerType=home"},{"key":"e_1_3_3_38_2","article-title":"Projective Transformations For Inter Plane Mappings","year":"1995","unstructured":"Projective_Transformations. 1995. Projective Transformations For Inter Plane Mappings. Retrieved from http:\/\/www.geom.uiuc.edu\/docs\/reference\/CRC-formulas\/node16.html. (1995).","journal-title":"http:\/\/www.geom.uiuc.edu\/docs\/reference\/CRC-formulas\/node16.html"},{"doi-asserted-by":"publisher","key":"e_1_3_3_39_2","DOI":"10.1145\/1098918.1098939"},{"doi-asserted-by":"publisher","unstructured":"Troels A. Rasmussen and Weidong Huang. 2019. SceneCam: Using AR to improve multi-camera remote collaboration(SA\u201919). Association for Computing Machinery New York NY USA 36\u201337. DOI:DOI:10.1145\/3355355.3361892","key":"e_1_3_3_40_2","DOI":"10.1145\/3355355.3361892"},{"key":"e_1_3_3_41_2","article-title":"Darknet: Open Source Neural Networks in C","author":"Redmon Joseph","year":"2013","unstructured":"Joseph Redmon. 2013\u20132016. Darknet: Open Source Neural Networks in C. Retrieved from http:\/\/pjreddie.com\/darknet\/. (2013\u20132016).","journal-title":"http:\/\/pjreddie.com\/darknet\/"},{"doi-asserted-by":"publisher","key":"e_1_3_3_42_2","DOI":"10.1109\/CVPR.2016.91"},{"unstructured":"Joseph Redmon and Ali Farhadi. 2018. YOLOv3: An incremental improvement. Retrieved from https:\/\/arxiv.org\/abs\/1804.02767","key":"e_1_3_3_43_2"},{"key":"e_1_3_3_44_2","article-title":"Sony IMX258 Camera Driver","year":"2022","unstructured":"Sony_IMX258_CAMdriver. 2022. Sony IMX258 Camera Driver. Retrieved from https:\/\/github.com\/raspberrypi\/linux\/blob\/rpi-5.4.y\/drivers\/media\/i2c\/imx219.c. (2022).","journal-title":"https:\/\/github.com\/raspberrypi\/linux\/blob\/rpi-5.4.y\/drivers\/media\/i2c\/imx219.c"},{"key":"e_1_3_3_45_2","article-title":"Mono-crystalline 9W Solar Panel","year":"2022","unstructured":"Sparkfun. 2022. Mono-crystalline 9W Solar Panel. Retrieved from https:\/\/www.sparkfun.com\/products\/13784. (2022).","journal-title":"https:\/\/www.sparkfun.com\/products\/13784"},{"key":"e_1_3_3_46_2","article-title":"SSD: Single-Shot MultiBox Detector Implementation in Keras","year":"2018","unstructured":"SSD. 2018. SSD: Single-Shot MultiBox Detector Implementation in Keras. Retrieved from https:\/\/github.com\/pierluigiferrari\/ssd_keras. (2018).","journal-title":"https:\/\/github.com\/pierluigiferrari\/ssd_keras"},{"key":"e_1_3_3_47_2","first-page":"1","article-title":"Towards scalable video analytics at the edge","author":"Stone Theodore","year":"2019","unstructured":"Theodore Stone, Nathaniel Stone, Puneet Jain, Yurong Jiang, Kyu-Han Kim, and Srihari Nelakuditi. 2019. Towards scalable video analytics at the edge. InProceedings of the 2019 16th Annual IEEE International Conference on Sensing, Communication, and Networking (SECON) (2019), 1\u20139.","journal-title":"Proceedings of the 2019 16th Annual IEEE International Conference on Sensing, Communication, and Networking (SECON)"},{"doi-asserted-by":"publisher","key":"e_1_3_3_48_2","DOI":"10.1109\/CVPR.2019.00900"},{"doi-asserted-by":"publisher","key":"e_1_3_3_49_2","DOI":"10.1109\/TCSVT.2015.2477935"},{"doi-asserted-by":"publisher","key":"e_1_3_3_50_2","DOI":"10.3390\/make5040083"},{"key":"e_1_3_3_51_2","article-title":"Tsai Camera Calibration","year":"2022","unstructured":"TSAI_Calib. 2022. Tsai Camera Calibration. Retrieved from https:\/\/homepages.inf.ed.ac.uk\/rbf\/CVonline\/LOCAL_COPIES\/DIAS1\/. (2022).","journal-title":"https:\/\/homepages.inf.ed.ac.uk\/rbf\/CVonline\/LOCAL_COPIES\/DIAS1\/"},{"doi-asserted-by":"publisher","key":"e_1_3_3_52_2","DOI":"10.1109\/JPROC.2010.2098830"},{"doi-asserted-by":"publisher","key":"e_1_3_3_53_2","DOI":"10.23919\/FUSION43075.2019.9011397"},{"doi-asserted-by":"publisher","key":"e_1_3_3_54_2","DOI":"10.1109\/PerCom53586.2022.9762398"},{"doi-asserted-by":"publisher","key":"e_1_3_3_55_2","DOI":"10.1145\/3386901.3388948"},{"doi-asserted-by":"publisher","key":"e_1_3_3_56_2","DOI":"10.1109\/CVPR.2016.461"},{"unstructured":"Yuanlu Xu Xiaobai Liu Lei Qin and Song-Chun Zhu. 2017. Cross-view people tracking by scene-centered spatio-temporal parsing(AAAI\u201917). AAAI Press.","key":"e_1_3_3_57_2"},{"doi-asserted-by":"publisher","key":"e_1_3_3_58_2","DOI":"10.1109\/INFOCOM48880.2022.9796984"},{"doi-asserted-by":"publisher","unstructured":"Juheon Yi Utku G\u00fcnay Acer Fahim Kawsar and Chulhong Min. 2025. Argus: Enabling cross-camera collaboration for video analytics on distributed smart cameras. IEEE Transactions on Mobile Computing 24 1 (2025) 117\u2013134. DOI:10.1109\/TMC.2024.3459409","key":"e_1_3_3_59_2","DOI":"10.1109\/TMC.2024.3459409"},{"doi-asserted-by":"publisher","key":"e_1_3_3_60_2","DOI":"10.1145\/3372224.3380881"},{"unstructured":"Alexey Bochkovskiy Chien-Yao Wang and Hong-Yuan Mark Liao. 2020. YOLOv4: Optimal speed and accuracy of object detection.","key":"e_1_3_3_61_2"},{"doi-asserted-by":"publisher","key":"e_1_3_3_62_2","DOI":"10.1007\/s11042-019-7204-5"}],"container-title":["ACM Transactions on Internet of Things"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3736420","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,7,15]],"date-time":"2025-07-15T11:08:12Z","timestamp":1752577692000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3736420"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,7,10]]},"references-count":61,"journal-issue":{"issue":"3","published-print":{"date-parts":[[2025,8,31]]}},"alternative-id":["10.1145\/3736420"],"URL":"https:\/\/doi.org\/10.1145\/3736420","relation":{},"ISSN":["2691-1914","2577-6207"],"issn-type":[{"type":"print","value":"2691-1914"},{"type":"electronic","value":"2577-6207"}],"subject":[],"published":{"date-parts":[[2025,7,10]]},"assertion":[{"value":"2024-03-05","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-04-11","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-07-10","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}