Chameleon: scalable adaptation of video analytics
The DOI matches its Crossref record; 24 of the 25 references it cites were found in a registry or an index.
Abstract
Applying deep convolutional neural networks (NN) to video data at scale poses a substantial systems challenge, as improving inference accuracy often requires a prohibitive cost in computational resources. While it is promising to balance resource and accuracy by selecting a suitable NN configuration (e.g., the resolution and frame rate of the input video), one must also address the significant dynamics of the NN configuration's impact on video analytics accuracy. We present Chameleon, a controller that dynamically picks the best configurations for existing NN-based video analytics pipelines. The key challenge in Chameleon is that in theory, adapting configurations frequently can reduce resource consumption with little degradation in accuracy, but searching a large space of configurations periodically incurs an overwhelming resource overhead that negates the gains of adaptation. The insight behind Chameleon is that the underlying characteristics (e.g., the velocity and sizes of objects) that affect the best configuration have enough temporal and spatial correlation to allow the search cost to be amortized over time and across multiple video feeds. For example, using the video feeds of five traffic cameras, we demonstrate that compared to a baseline that picks a single optimal configuration offline, Chameleon can achieve 20-50% higher accuracy with the same amount of resources, or achieve the same accuracy with only 30--50% of the resources (a 2-3X speedup).
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 1×
“, [17, 18]) which have led to a continuous stream of increasingly accurate models for object detection and classification.”
BackgroundHe, K., Gkioxari, G., Dollar, P., & Girshick, R. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.322 · Cited by 29,137
Girshick, R. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.169 · Cited by 28,105
The pascal visual object classes (VOC) challenge
Everingham, M., Van Gool, L., Williams, C. K. I., Winn, J., & Zisserman, A. (2010) · International Journal of Computer Vision · doi:10.1007/s11263-009-0275-4 · Cited by 19,599
MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
Howard, A. G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., Andreetto, M., & Adam, H. (2017) · arXiv · doi:10.48550/arxiv.1704.04861 · Cited by 19,015
Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning
Szegedy, C., Ioffe, S., Vanhoucke, V., & Alemi, A. (2017) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v31i1.11231 · Cited by 4,636
Speed/Accuracy Trade-Offs for Modern Convolutional Object Detectors
Huang, J., Rathod, V., Sun, C., Zhu, M., Korattikara, A., Fathi, A., Fischer, I., Wojna, Z., Song, Y., Guadarrama, S., et al. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.351 · Cited by 2,743 · cited 1×
“The inference models are implemented in Tensorflow and are pretrained on standard image datasets [23], and the switching of video frame rate and resolution is done by FFmpeg [10].”
Method(1993) · Choice Reviews Online · doi:10.5860/choice.31-0986 · Cited by 1,540
Analysis of Thompson Sampling for the multi-armed bandit problem
Agrawal, S., & Goyal, N. (2011) · arXiv (Cornell University) · doi:10.48550/arxiv.1111.1797 · Cited by 819 · cited 1×
“al [19] use Thompson sampling [11] to optimize the layout of a multi-component web page; they use greedy hill climbing to select the next layout, similar to Chameleon, but Chameleon exploits more independence structure and monotonicity.”
MethodReal-Time Video Analytics: The Killer App for Edge Computing
Ananthanarayanan, G., Bahl, P., Bodik, P., Chintalapudi, K., Philipose, M., Ravindranath, L., & Sinha, S. (2017) · Computer · doi:10.1109/mc.2017.3641638 · Cited by 496 · cited 2×
“As we move to the scenario of edge camera analytics [13], such a separation would be hard to achieve.”
BackgroundSelf-citation
Research works printed but not linked: 2
- FFmpeg. http://ffmpeg.org/ , 2000 --2018. FFmpeg. http://ffmpeg.org/, 2000--2018.
- W. E. Hoover and M. Rockville . Algorithms for confidence circles and ellipses . Citeseer , 1984 . W. E. Hoover and M. Rockville. Algorithms for confidence circles and ellipses. Citeseer, 1984.
Cited by
Items 301–400 of 575
LENS: Bandwidth-efficient video analytics with adaptive super resolution
Shan, M., Zhang, S., Xiao, M., & Zhao, Y. (2022) · Computer Networks · doi:10.1016/j.comnet.2022.109392 · Cited by 6
AI Workload Allocation Methods for Edge-Cloud Computing: A Review
Rafea, S. A., & Jasim, A. D. (2023) · Al-Iraqia Journal of Scientific Engineering Research · doi:10.58564/ijser.2.4.2023.125 · Cited by 6
Contention aware DRAM caching for CXL-enabled pooled memory
Tirumalasetty, C., & Reddy, N. (2024) · International Symposium on Memory Systems (MEMSYS) · doi:10.1145/3695794.3695808 · Cited by 6
“ML techniques like large language models(LLM) [22, 39, 69, 74, 75], video processing [32, 37, 76] and others, rely on large amounts of data for training and some-times for retrieval.”
Dependable machine intelligence at the tactical edge
Misra, A., Jayarajah, K., Weerakoon, D., Tandriansyah, R., Yao, S., & Abdelzaher, T. (2019) · Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications · doi:10.1117/12.2522656 · Cited by 6
Small, Accurate, and Fast Re-ID on the Edge: The SAFR Approach
Suprem, A., Pu, C., & Ferreira, J. E. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-59824-2_5 · Cited by 6
Scaling Video Analytics Systems to Large Camera DeploymentsPreprint
Jain, S., Ananthanarayanan, G., Jiang, J., Shu, Y., & Gonzalez, J. E. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1809.02318 · Cited by 7
Self-citationA Survey of Performance Optimization in Neural Network-Based Video Analytics SystemsPreprint
Ibrahim, N. B., Maurya, P., Jafari, O., & Nagarkar, P. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2105.14195 · Cited by 6
“Chameleon [14] is a video analytics system that reduces profiling costs, optimizes resource consumption, and improves inference accuracy of video analytics pipelines.”
BackgroundReXCam: Resource-Efficient, Cross-Camera Video Analytics at Scale
Jain, S., Zhang, X., Zhou, Y., Ananthanarayanan, G., Jiang, J., Shu, Y., & Gonzalez, J. (2018) · arXiv · doi:10.48550/arxiv.1811.01268 · Cited by 5
Self-citationA Resource-Efficient Multiple Recognition Services Framework for IoT Devices
Ding, C., Liu, Z., Zhou, A., Yu, J., Li, Y., & Wang, S. (2024) · IEEE Transactions on Services Computing · doi:10.1109/tsc.2024.3512949 · Cited by 5
“A large number of excellent cloud-based training neural network models have sprung up [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24].”
Rawlley, O., Gupta, S. R., Mahajan, K., Shrivastava, A., & Jain, E. (2025) · Future Generation Computer Systems · doi:10.1016/j.future.2025.107817 · Cited by 5
Flex: Fast, Accurate DNN Inference on Low-Cost Edges Using Heterogeneous Accelerator Execution
Sen, T., Shen, H., & Iyer, A. P. (2025) · European Conference on Computer Systems · doi:10.1145/3689031.3696067 · Cited by 5
“5 , 5 ] seconds and their accuracy requirement over 100 inputs was randomly chosen from [ 80% , 95% ] [23, 38], We empirically set 𝛼 = 0 .”
ViEdge: Video Analytics on Distributed Edge
Hou, X., Guan, Y., & Han, T. (2025) · ACM Transactions on Internet of Things · doi:10.1145/3742794 · Cited by 5
Mitigating Tail Latency for On-Device Inference With Load-Balanced Heterogeneous Models
Yuan, M., Zhang, L., Duan, D., Zeng, L., Song, M.-H., Li, Z., Xing, G., & Li, X.-Y. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3588430 · Cited by 5
“(cid:2) Evaluations of InferRouter on three representative on-device inference tasks show that InferRouter signifi-cantly outperforms complementary state-of-the-art approaches [16], [19], [20] in both accuracy and P99 latency metrics.”
InfluentialECORE: Energy-Conscious Optimized Routing for Deep Learning Models at the Edge
Alqahtani, D. K., Rodriguez, M. A., Cheema, M. A., Rezatofighi, H., & Toosi, A. N. (2026) · IEEE International Conference on Pervasive Computing and Communications · doi:10.1109/percom67906.2026.11524531 · Cited by 5
MVTEE: Multi-Variant Trusted Execution for Secure Model Inference
Qin, K., & Gu, D. (2025) · International Middleware Conference · doi:10.1145/3721462.3730956 · Cited by 5
“ML inference has massive adoption in applications [48, 49].”
ST-360: Spatial–Temporal Filtering-Based Low-Latency 360-Degree Video Analytics Framework
Li, J., Liao, J., Chen, B., Nguyen, A., Tiwari, A., Zhou, Q., Yan, Z., & Nahrstedt, K. (2024) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3694685 · Cited by 5
“Yu et al. [73] and Wang et al. [60] each proposed their unique 360-degree object detectors, utilizing R-CNN and S-CNN, respectively, to directly process 360-degree images in ERP format.”
FastTuner: Fast Resolution and Model Tuning for Multi-Object Tracking in Edge Video Analytics
Xu, R., Nalaie, K., & Zheng, R. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3526573 · Cited by 5
“In the inference stage, FastTuner runs the most expensive configuration, also known as the golden configuration [7] every K frames to determine the optimal con-figuration based on the heatmaps produced by the detectability branch.”
InfluentialA Video Capturing and Processing Platform Based on Mobile Edge Computing
Zhao, Y., An, R., Ou, D., & Jiang, C. (2020) · International Conference on Computer, Information and Telecommunication Systems · doi:10.1109/cits49457.2020.9232607 · Cited by 5
“Video streaming and analytics from these widely deployed cameras are becoming more and more popular, which enables the business intelligence mining, intelligent transportation optimization, as well as healthcare improvements and lift security[1,2,3,4].”
BackgroundControlled Intentional Degradation in Analytical Video Systems
He, W., & Cafarella, M. (2022) · ACM SIGMOD International Conference on Management of Data (SIGMOD) · doi:10.1145/3514221.3517899 · Cited by 5
EC-MASS: Towards an efficient edge computing-based multi-video scheduling system
Yang, S., Dong, Q., Cui, L., Chen, X., Lei, S., Wu, Y., & Luo, C. (2022) · Computer Communications · doi:10.1016/j.comcom.2022.07.002 · Cited by 5
Towards Exploiting Stakeholder Resources for Data-Driven Resilient Hazard Management
Fu, X., Yuan, T., Wang, W., Wolk, F., & Strunz, G. (2022) · IEEE International Conference on Distributed Computing Systems Workshops (ICDCSW) · doi:10.1109/icdcsw56584.2022.00055 · Cited by 5
“A typical video analytics system follows this pipeline [28], [29].”
BackgroundCreating Robust Deep Neural Networks with Coded Distributed Computing for IoT
Hadidi, R., Cao, J., Asgari, B., & Kim, H. (2023) · IEEE International Conference on Edge Computing and Communications (EDGE) · doi:10.1109/edge60047.2023.00029 · Cited by 5
Optimizing Video Selection LIMIT Queries with Commonsense Knowledge
He, W., Sabek, I., Lou, Y., & Cafarella, M. (2024) · Proceedings of the VLDB Endowment · doi:10.14778/3654621.3654639 · Cited by 5
Reinforcement Learning based Orchestration for Elastic Services
Argerich, M. F., Cheng, B., & Fürst, J. (2019) · IEEE World Forum on Internet of Things (WF-IoT) · doi:10.1109/wf-iot.2019.8767180 · Cited by 5
“In Chameleon [8], the performance of video analytics applications is optimized by performing automatic adaptation of its configurations.”
Dataset Culling: Towards Efficient Training of Distillation-Based Domain Specific Models
Yoshioka, K., Lee, E. H., Wong, S. S., & Horowitz, M. (2019) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2019.8803462 · Cited by 5
Why Is the Video Analytics Accuracy Fluctuating, and What Can We Do About It?
Paul, S., Rao, K., Coviello, G., Sankaradas, M., Po, O., Hu, Y., & Chakradhar, S. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-25056-9_28 · Cited by 5
Crowd2: Multi-agent Bandit-based Dispatch for Video Analytics upon Crowdsourcing
Chen, Y., Zhang, S., Yan, Y., Jin, Y., Chen, N., Ji, M., & Xiao, M. (2023) · IEEE Conference on Computer Communications (INFOCOM) · doi:10.1109/infocom53939.2023.10228927 · Cited by 5
Yadav, P. (2021) · Arrow@dit (Dublin Institute of Technology) · doi:10.13025/17114 · Cited by 5
A Review of the Convergence of 5G/6G Architecture and Deep LearningPreprint
Odeyomi, O. T., Akintade, O. O., Olowu, T. O., & Zaruba, G. V. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2208.07643 · Cited by 5
“Some of the works on resource sharing are VideoStorm[115], Chameleon [116], Mainstream [117].”
BackgroundHydro: Adaptive Query Processing of ML QueriesPreprint
Kakkar, G. T., Cao, J., Sengupta, A., Arulraj, J. J. P., & Kim, H. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2403.14902 · Cited by 5
“Similarly, Chameleon [27] selects the most optimal algorithm through sliding window-based profiling.”
MethodGuan, X., Zhang, Z., Lin, Z., Sun, Z., Duan, T., Fang, Z., Wang, R., Cui, H., Ni, W., Luo, J., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2605.06027 · Cited by 5
“Note that several other techniques also reduce the cost of video analytics, including ROI filtering [36], [37], input resolution adaptation [38], [39], frame sampling that skips inference on selected frames [36], [37], and model compression via quantization or pruning [40], [41].”
Cheng, S., Wang, Z., Feng, F., Zhang, Y., Bi, T., & Jiang, T. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3581523 · Cited by 4
“Jiang et al. [12] introduced Chameleon, a framework that dynamically searches for optimal configurations of video analysis pipelines, offering efficient and scalable adaptive video analysis.”
DeVA: An Edge-Assisted Video Analytics Framework for Depth Estimation
Chen, S., Yin, J., Zhong, R., & Liu, F. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3588864 · Cited by 4
“For example, some approaches compress the resolution or filter offloaded frames to reduce the size of offloaded video [6], [7], [8].”
Real-Time Video Analytics for Urban Safety: Deployment over Edge and End Devices
Ghasemi, M., Fu, Y., Ouyang, X., Wang, P., Türkcan, M. K., Tavori, J., Kleisarchaki, S., Calmant, T., Gürgen, L., Kostić, Z., et al. (2025) · ACM/IEEE Symposium on Edge Computing (SEC) · doi:10.1145/3769102.3770618 · Cited by 4
Huang, Y., Zharfan, F., Hendrawan, H., Gunawi, H. S., & Jiang, J. (2024) · IEEE International Conference on Edge Computing and Communications (EDGE) · doi:10.1109/edge62653.2024.00017 · Cited by 4
“…quickly adapt these configurations, including offline profiling ( e.g. [29]) or periodically reprofiling ( e.g. [30]) the relationship between the configuration settings and its inference accuracy, and cheap heuristics to derive a good setting of configuration from input video content ( e.g. [12]).”
BackgroundInfluentialSelf-citationScrava: Super Resolution-Based Bandwidth-Efficient Cross-Camera Video Analytics
Liang, Y., Zhang, S., & Wu, J. (2024) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2024.3461879 · Cited by 4
“…as MOTA is a multi-object tracking metric at the object detection level, which calculates the bijection between the predicted bounding box set prDets and the ground truth bounding box set gtDets in each frame, and two bounding boxes are considered as a match ( TP ) if they are spatially similar…”
Tu, J., Xu, Q., & Chen, C. (2021) · Annual Conference of the IEEE Industrial Electronics Society (IECON) · doi:10.1109/iecon48115.2021.9589140 · Cited by 4
AdaConfigure: Reinforcement Learning-Based Adaptive Configuration for Video Analytics Services
He, Z., Wang, Y., Tang, C., Wang, Z., Zhu, W., Guo, C., & Chen, Z. (2022) · Lecture notes in computer science · doi:10.1007/978-3-030-98358-1_20 · Cited by 4
Resource and Bandwidth-Aware Video Analytics with Adaptive Offloading
Zhang, L., Zhong, Y., Liu, J., & Cui, L. (2023) · IEEE International Conference on Mobile Ad-Hoc and Sensor Systems (MASS) · doi:10.1109/mass58611.2023.00021 · Cited by 4
“Other works propose to profile the trade-off between accuracy and bandwidth and use the profiling result to adjust transmission configurations such as data rate [14], resolution & frame rate [15], and video batch size [16].”
BackgroundRaven: Vision-Based Connected Vehicle Safety Platform Using Infrastructure Sensing, 5G, and MEC
Ali, K., Yu, B., Kumar, V. V., Krishnan, H., Bai, F., Rehan, M., & Kumar, A. V. (2024) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2024.3430815 · Cited by 4
“Video Streaming and Analytics: Significant work has been expended to improve the efficiency of video analytics pipelines [11], [30], [31], [32], [33], [34], [35], [36], [37], [38].”
BackgroundLeveraging Multi-Modal Data for Efficient Edge Inference Serving
Wolfrath, J., Achanta, A., & Chandra, A. (2024) · IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGrid) · doi:10.1109/ccgrid59990.2024.00053 · Cited by 4
“Chameleon attempts to dynamically identify optimal configurations for each stream to improve efficiency [40].”
BackgroundLyu, Y., Liu, L., Wang, X., Fan, Z., Wang, J., & Gao, G. (2025) · Machine Learning and Knowledge Extraction · doi:10.3390/make7040117 · Cited by 4
“VideoStorm [31] and Chameleon [32] adopted frame rate tuning resolution resizing to reduce the transmitted video volume.”
Optimizing Machine Learning Inference Queries for Multiple Objectives
Li, Z., Schönfeld, M., Hai, R., Bozzon, A., & Katsifodimos, A. (2023) · IEEE International Conference on Data Engineering Workshops (ICDEW) · doi:10.1109/icdew58674.2023.00017 · Cited by 4
Rethinking Wireless Network Management Through Sensor-driven Contextual Analysis
Irshad, S., Rozner, E., Bhartia, A., & Chen, B. (2020) · IEEE Workshop on Mobile Computing Systems and Applications (WMCSA) · doi:10.1145/3376897.3377863 · Cited by 4
DAEMON Deliverable 3.1: Initial design of real-time control and VNF intelligence mechanisms
Gramaglia, M., Mhaisen, N., García‐Saavedra, A., Garcia, G., Chang, C., Vleeschauwer, D. D., Molina, M., Slamnik–Kriještorac, N., Soto, P., Camelo, M., et al. (2021) · Zenodo (CERN European Organization for Nuclear Research) · doi:10.5281/zenodo.5745432 · Cited by 4
Zhang, E., Daum, M., He, D., Haynes, B. M., Krishna, R., & Bałazińska, M. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2301.00929 · Cited by 4
“Many recent VDBMSs have been proposed and focus on a wide range of data management challenges, including fast inference over videos [3, 5, 34, 44, 47], storage optimization [18, 25, 64], efficient dataflow processing [49], preprocessing and indexing [6, 28, 36], exploration and organization [19], privacy [9], and tuni…”
BackgroundMondrian: On-Device High-Performance Video Analytics with Compressive Packed InferencePreprint
Jeon, C., Kim, S., Yi, J., & Lee, Y. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2403.07598 · Cited by 4
“Recent works proposed profiling techniques to determine the optimal video quality (e.g., resolution or quantization parameter) for live video analytics pipeline by running the DNN with different parameters repetitively on the cloud [24], [34].”
EdgeSync: Faster Edge-model Updating via Adaptive Continuous Learning for Video Data DriftPreprint
Zhao, P. B., Dong, R., Wang, G., & Zhao, C. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2406.03001 · Cited by 4
“To balance accuracy and speed, Chameleon[20] designs a controller to dynamically select parameters in the video analysis system; Wang et al. [39] addresses the joint configuration tuning problem for video parameters and server-less computation resources, they propose an algorithm utilizing Markov approximation to sele…”
MethodSatisfying service level objectives in stream processing systems
Kalim, F. (2020) · Illinois Digital Environment for Access to Learning and Scholarship (University of Illinois at Urbana-Champaign) · Cited by 4
FireLoc: Low-latency Multi-modal Wildfire Geolocation
Fu, X., Hu, Y. X., Sutrave, P., Beerel, P. A., & Raghavan, B. (2024) · ACM Conference on Embedded Networked Sensor Systems (SenSys) · doi:10.1145/3666025.3699318 · Cited by 3
RA-MOSAIC : Resource Adaptive Edge AI Optimization over Spatially Multiplexed Video Streams
Gokarn, I., Hu, Y., Abdelzaher, T., & Misra, A. (2025) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3715133 · Cited by 3
“…2022a; Zhang et al., 2021], content masking [Liu et al., 2022c], input filtering [Chen et al., 2015], micro classifiers and filters at the camera [Jiang et al., 2018], and multi-resolution frame optimization and transmission [Jiang et al., 2021; Wu et al., 2022], all optimize the content of the…”
CamTuner: Adaptive Video Analytics Pipelines via Real-Time Automated Camera Parameter Tuning
Paul, S., Rao, K., Coviello, G., Sankaradas, M., Hu, Y., & Chakradhar, S. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3540667 · Cited by 3
“…data transformations such as adjustments in compression quality, resolution, or frame appearance (e.g., brightness, contrast etc.) on individual frames of the video feed captured by the camera [27], [56], and 3) tuning camera parameters in order to help the camera capture good quality videos.”
ARC: Approximate Relevant Clip Query in Large-Scale Video Repositories
Chen, Y., Jing, Y., Yu, Z., Yu, X., He, Z., Zhang, K., & Wang, X. S. (2025) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3726302.3729896 · Cited by 3
“Recent research involves programming interfaces [29, 32], query optimization [1, 2, 4, 45], and execution scheduling [12, 19, 32, 35, 43].”
Scaling bitrate in video compression using tuned parameters for efficient video analytics
Masykuroh, K., Hendrawan, H., & Mulyana, E. (2025) · Telematics and Informatics Reports · doi:10.1016/j.teler.2025.100274 · Cited by 3
Zuo, L., Zheng, Q., Yang, Z., & Ji, W. (2025) · IEEE International Conference on Parallel and Distributed Systems (ICPADS) · doi:10.1109/icpads67057.2025.11323112 · Cited by 3
Towards resource-efficient detection-driven processing of multi-stream videos
Arefeen, M. A., & Uddin, M. Y. S. (2021) · ACM Annual International Conference on Mobile Computing and Networking (MobiCom) · doi:10.1145/3447993.3482866 · Cited by 3
“In [3], a configuration profile is generated based on different frame sizes and model sizes.”
MethodYou Only Look Once in Panorama: Object Detection for 360° Videos with MLaaS
Shen, L., Zhang, M., Zhang, C., & Liu, J. (2024) · International Workshop on Network and Operating Systems Support for Digital Audio and Video (NOSSDAV) · doi:10.1145/3651863.3651876 · Cited by 3
Beimin: Serverless-based Adaptive Real-Time Video Processing
Zhang, J., Meng, Z., & Xu, M. (2024) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme57554.2024.10687715 · Cited by 3
Distributed No-Regret Learning for Multi-Stage Systems with End-to-End Bandit Feedback
Hou, I.-H. (2024) · ACM International Symposium on Theory, Algorithmic Foundations, and Protocol Design for Mobile Networks and Mobile Computing (MobiHoc) · doi:10.1145/3641512.3686369 · Cited by 3
Zhang, Y., Zhang, W., Yuan, M., Xu, L., Yan, C., Gong, T., & Du, H. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3526359 · Cited by 3
“It can be categorized into two types: (i) rule-based algorithms [8], [12], [13], [15], [16], [20], and (ii) learning-based algorithms [7].”
Leveraging DVFS for Energy-Efficient and QoS-aware Edge Video Analytics
Loudaros, I., Zervakis, G., Soudris, D., & Masouros, D. (2025) · doi:10.1145/3721889.3721925 · Cited by 3
SneakPeek: Data-Aware Model Selection and Scheduling for Inference Serving on the Edge
Wolfrath, J., Frink, D., & Chandra, A. (2025) · ACM Symposium on Cloud Computing (SoCC) · doi:10.1145/3772052.3772217 · Cited by 3
“For all approaches, we observe an increase in scheduling overhead as the number of requests increases.”
Adaptive Scheduling for Edge-Assisted DNN Serving
He, J., Yang, C., He, Z., Baig, G., & Qiu, L. (2023) · IEEE International Conference on Mobile Ad-Hoc and Sensor Systems (MASS) · doi:10.1109/mass58611.2023.00071 · Cited by 3
“Existing approaches, such as Deepdecision [50], DARE [36], Chameleon [30]), optimize latency by adaptively adjusting the video resolution and frame rate according to the network condition and server workload.”
MethodKossoski, C., Simão, J. M., & Lopes, H. S. (2024) · Applied Sciences · doi:10.3390/app14093566 · Cited by 3
“Specifically in the area of video queries, many authors have used GPUs to perform data ingestion, preprocessing, and object labeling [3–5, 30– 33 ], among Besides, it is possible to use Field-Programmable Gate Arrays (FPGA) for deep learning tasks instead of GPUs [36].”
BackgroundMethodPerformance-cost trade-offs in service orchestration for edge computing
Balouek‐Thomert, D. (2024) · International Conference on Scientific and Statistical Database Management · doi:10.1145/3676288.3676307 · Cited by 3
“Online/offline configurations adaptation is currently a promising solution to address the issue of limited resources [9, 10, 12, 15, 17].”
Towards Dynamically Reconfigurable IoT Camera Virtualization for Video Analytics Edge Cloud Services
Lee, D., Jang, S. Y., Shin, B., & Lee, Y. (2019) · IEEE Internet Computing · doi:10.1109/mic.2019.2893872 · Cited by 3
Robust, Extensible, and Fast: Teamed Classifiers for Vehicle Tracking in Multi-Camera Networks
Suprem, A., Lima, R. A., Padilha, B., Ferreira, J. E., & Pu, C. (2019) · IEEE International Conference on Cognitive Machine Intelligence (CogMI) · doi:10.1109/cogmi48466.2019.00013 · Cited by 3
“Index Terms—teamed classifiers, video analytics, conditional computation, vehicle re-identification, convolutional attention, targeted attention I. INTRODUCTION The ubiquity of large-scale camera networks have coincided with the emergence of real-time video management platforms such as Chameleon [1], Kestrel [2], Vide…”
BackgroundMethodInfluentialSpider: Next generation Live Video Analytics over Millimeter-Wave Networks
Li, Z., Shu, Y., Ananthanarayanan, G., Shangguan, L., Jamieson, K., & Bahl, V. (2020) · Cited by 3
Self-citationOptimal Offloading of Computing-intensive Tasks for Edge-aided Maritime UAV Systems
Li, H., Wu, S., Li, D., Jiao, J., Zhang, N., & Zhang, Q. (2022) · IEEE Vehicular Technology Conference · doi:10.1109/vtc2022-spring54318.2022.9860881 · Cited by 3
“Moreover, the performance of target detection algorithm under deep learning is often positively correlated with the image resolution [10], and the current network model often cannot achieve 100% average precision (AP), which further affects the performance of the tracking algorithm [11].”
BackgroundRETRACTED: TRUST: A Novel Framework for Vehicle Trajectory Recovery from Urban-Scale Videos
Ji, W., & Chen, X. (2022) · Sensors · doi:10.3390/s22249948 · Cited by 3
“A large number of video database systems have emerged, such as NoScope [3], Focus [4], Chameleon [5], BlazeIt [6], and MIRIS [7], which can efficiently deal with all types of video queries in target detection, identification, and tracking tasks.”
BackgroundRetractedMLink: Linking Black-Box Models From Multiple Domains for Collaborative Inference
Yuan, M., Zhang, L., Zheng, Z., Zhang, Y.-N., & Li, X. (2023) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3283780 · Cited by 3
“Adaptive configuration [17] and multi-model scheduling [18] were proposed to make inference workloads adaptive to the dynamics of input content.”
BackgroundOptimizing ML Inference Queries Under Constraints
Li, Z., Schönfeld, M., Sun, W., Fragkoulis, M., Hai, R., Bozzon, A., & Katsifodimos, A. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-34444-2_4 · Cited by 3
“Machine learning (ML) is increasingly used across application domains such as video analytics [24, 58], autonomous driving [56], content moderation [16], traffic monitoring [41] and crowd detection [27].”
BackgroundOSCA: Online User-managed Server Selection and Configuration Adaptation for Interactive MAR
Shi, X., Zhang, S., Cheng, K., Chen, Y., Zhu, A., & Lu, S. (2023) · International Workshop on Quality of Service · doi:10.1109/iwqos57198.2023.10188796 · Cited by 3
EdgeMA: Model Adaptation System for Real-Time Video Analytics on Edge Devices
Wang, L., Zhang, N., Qu, X., Wang, J., Wan, J., Li, G., Hu, K., Jiang, G., & Xiao, J. (2023) · Lecture notes in computer science · doi:10.1007/978-981-99-8079-6_23 · Cited by 3
Cost-Efficient Cloud-Edge Video Analytics with Hybird IaaS and FaaS Resources
Peng, J., Zhou, Z., Zhao, K., Ma, H., & Chen, X. (2023) · IEEE International Conference on Parallel and Distributed Systems (ICPADS) · doi:10.1109/icpads60453.2023.00335 · Cited by 3
Accelerating Low-Cost Edge-Based Real-Time Video Analytics using Task Scheduling
Ong, K. S. H., Niyato, D. T., So–In, C., & Friedrichs, T. (2021) · IEEE World Forum on Internet of Things (WF-IoT) · doi:10.1109/wf-iot51360.2021.9595588 · Cited by 3
MOTrack: Real-time Configuration Adaptation for Video Analytics through Movement Tracking
Wu, F., Gao, L., Zhou, T., & Wang, X. (2021) · Globecom · doi:10.1109/globecom46510.2021.9685159 · Cited by 3
“The ground truth for measuring video analytics accuracy is based on the most expensive configuration with the highest frame rate and resolution [2].”
BackgroundMethodInfluentialAdaptive Cross-Camera Video Analytics at the Edge
Chen, K., Zhu, Y., Han, Z., & Wang, X. (2022) · IEEE International Conference on Mobile Ad-Hoc and Sensor Systems (MASS) · doi:10.1109/mass56207.2022.00063 · Cited by 3
Suprem, A., Vaidya, S., Cherkadi, S., Singh, P., Ferreira, J. E., & Pu, C. (2022) · IEEE International Conference on Cognitive Machine Intelligence (CogMI) · doi:10.1109/cogmi56440.2022.00021 · Cited by 3
“This is addressed with an object detector such as YOLO or Mask-RCNN, and in case of video streams, a detector UDF over a query engine, similar to [23].”
MethodExploiting mobility for predictive urban analytics & operations
Jayarajah, K. (2019) · Singapore Management University Institutional Knowledge (InK) (Singapore Management University) · Cited by 3
Boggart: Accelerating Retrospective Video Analytics via Model-Agnostic Ingest Processing.Preprint
Agarwal, N., & Netravali, R. (2021) · arXiv (Cornell University) · Cited by 3
Cerberus: Real-Time Video Anomaly Detection via Cascaded Vision-Language ModelsPreprint
Zheng, Y., Shi, X., Chen, J., & Shu, Y. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.16290 · Cited by 3
“Traditional video analytics systems require complex and careful query planning, which requires extensive domain-specific experience [7, 15, 61, 63].”
TagTeam: Towards wearable-assisted, implicit guidance for human-drone teams
Jayarajah, K., Gangopadhyay, A., & Waytowich, N. R. (2022) · doi:10.1145/3556560.3560715 · Cited by 2
“Several recent works have explored optimization techniques for networked sensors to achieve efficient querying [11, 15].”
BackgroundAgile3D: Adaptive Contention- and Content-Aware 3D Object Detection for Embedded GPUs
Wang, P., Liu, Z., Bagchi, S., Xu, R., Bagchi, S., Li, Y., & Chaterji, S. (2025) · International Conference on Mobile Systems, Applications, and Services · doi:10.1145/3711875.3729147 · Cited by 2
Uirapuru: Timely Video Analytics for High-Resolution Steerable Cameras on Edge Devices
Apostolo, G. H., Bauszat, P., Nigade, V., Bal, H. E., & Wang, L. (2025) · ACM Annual International Conference on Mobile Computing and Networking (MobiCom) · doi:10.1145/3680207.3765260 · Cited by 2
SoK: Evaluation of Methods for Privacy Preserving Edge Video Analytics
Joseph, A., & Ganapathy, V. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-13714-2_24 · Cited by 2
CIVS: Communication-Aware Industrial Video Surveillance With Edge-End Collaboration
Tu, J., Lu, Y., Yan, Y., Chen, C., & Guan, X. (2026) · IEEE Internet of Things Journal · doi:10.1109/jiot.2026.3654565 · Cited by 2
“From the perspective of edge-enabled industrial video processing, another solution is to devise resource-scheduling algorithms [1], [22], [23], [24] that meet the low-latency requirements of industrial video surveillance on a single edge node.”
InfluentialBudget-Adaptive Routing: Skipping the Weak When the Strong Answers Anyway
Geng, W., Mohan, N., & Ott, J. (2026) · ACM SIGCOMM Conference (SIGCOMM) · doi:10.1145/3789240.3828740 · Cited by 2
MACEdge: Real-Time Video Analytics Based on Multi-Access Collaborative Edge Computing
Zhong, D., Zhu, M., & Huang, B. (2023) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc53992.2023.10393957 · Cited by 2
“Prior research has shown that the relationship between resolution and accuracy can be modeled as a concave function [4], [14].”
BackgroundA Contributory Public-Event Recording and Querying System
Joseph, A., Yadav, N., Ganapathy, V., & Behl, D. (2023) · ACM/IEEE Symposium on Edge Computing (SEC) · doi:10.1145/3583740.3628445 · Cited by 2
OVIDA: Orchestrator for Video Analytics on Disaggregated Architecture
Singh, M., Rachuri, S. P., Cao, B. B., Sharma, A., Bhumireddy, V., Bronzino, F., Das, S. R., Gandhi, A., & Jain, S. (2024) · IEEE/ACM Symposium on Edge Computing (SEC) · doi:10.1109/sec62691.2024.00019 · Cited by 2
“First, they require all or multiple models to be pre-loaded into the memory [14], [17]–[20].”
InfluentialSpliceosome: On-Camera Video Thinning and Tuning for Timely and Accurate Analytics
Chen, N., Zhang, S., Wu, J., Huang, H., & Lu, S. (2025) · IEEE Transactions on Networking · doi:10.1109/ton.2025.3526218 · Cited by 2
“…to numerous research efforts focused on different areas such as camera-edge-cloud collaboration [18], [35], [36], [37], [38], [39], [40], [41], [42], [57], DNN sharing [44], [45], [46], [47], and finding the optimal balance between accuracy and cost [17], [43], [48], [49], [50], [51], [52], [53].”
An On-Device Deep Learning Approach to Battery Saving on Industrial Mobile Terminals
Choi, I., & Kim, H. (2020) · Sensors · doi:10.3390/s20144044 · Cited by 2
“Although our work is not about techniques for optimizing deep learning on mobile devices, we note that many have been developed for that purpose, such as model compression [7], model partition [8], model early-exit [9], edge caching [10], input filtering [11], model selection [12], support for multitenancy [13], and a…”
Edge-Cloud Collaboration for Human Activity Recognition on Multiple Subjects
Xiao, W., Xie, L., Ning, J., Fu, Z., Zhao, M., Lin, Z., & Lin, Q. (2022) · IEEE International Symposium on a World of Wireless, Mobile and Multimedia Networks (WoWMoM) · doi:10.1109/wowmom54355.2022.00030 · Cited by 2
Data Exchange Mechanism for Real-Time Object Detection in Cloud-Edge IoT System
Miao, W., Zeng, Z., Huang, J. L., Li, S., Xia, Y., Bi, S., & Wang, X. (2023) · Wireless Communications and Mobile Computing · doi:10.1155/2023/3842738 · Cited by 2
“Chameleon [7] dynamically selects the optimal resource allocation for the existing deep neural network based video analysis pipeline.”
BackgroundElixir: A System to Enhance Data Quality for Multiple Analytics on a Video Stream
Paul, S., Rao, K., Coviello, G., Sankaradas, M., Hu, Y., & Chakradhar, S. (2023) · IEEE International Conference on Smart Computing · doi:10.1109/smartcomp58114.2023.00030 · Cited by 2
“Several works e.g., Chameleon [14], Videostorm [15], and AWStream [16] adapt camera settings like frame resolution and frame per second (fps) while processing single AU.”
BackgroundTowards Timely Edge-assisted Video Analytics Services
Li, X., Zhang, S., Huang, Y., Ma, X., Wang, Z., & Luo, H. (2023) · IEEE International Conference on Web Services (ICWS) · doi:10.1109/icws60048.2023.00063 · Cited by 2
“Video analytics is a critical component in modern smart city and intelligent transportation systems, which requires delivering accurate results to users timely [11].”
BackgroundCAC: Content-aware Adaptive Configuration for Edge Video Analytics
Cao, Z. (2023) · IEEE International Conference on Parallel and Distributed Systems (ICPADS) · doi:10.1109/icpads60453.2023.00347 · Cited by 2
Volatile MAB-based Configuration Selection for Offloading Video Analytics Tasks to Edges
Liang, Y., Zhang, S., & Wu, J. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10890608 · Cited by 2
BlockHybrid: Accelerating Object Detection Pipelines With Hybrid Block-Wise Execution
Xu, R., Nalaie, K., & Zheng, R. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3554167 · Cited by 2
“However, this problem can be solved by offline or online profiling [42], which we leave for future endeavor.”
Metadata
- Authors
- Jiang, J., Ananthanarayanan, G., Bodik, P., Sen, S., & Stoica, I.
- Affiliations
- University of Chicago (US)
- Microsoft (United States) (US)
- Microsoft Research
- University of California, Berkeley (US)
- Venue
- Proceedings of the 2018 Conference of the ACM Special Interest Group on Data Communication
- Publisher
- ACM
- Published
- 2018-08-07
- Pages
- 253-266
- OpenAlex
- W2887117815
- Cited by
- 467
- References
- 25
- Citations by year
- 2026: 33 · 2025: 58 · 2024: 69 · 2023: 74 · 2022: 70 · 2021: 69 · 2020: 50 · 2019: 38 · 2018: 6
- FWCI
- 12.28
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- ACM SIGCOMM Conference (SIGCOMM) (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Video Surveillance and Tracking Methods
- Physical Sciences › Computer Science › Information Systems › Advanced Database Systems and Queries
- Keywords
- video analysis, deep convolutional neural networks, object detection, chameleons
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.