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.”
MethodErnest: efficient performance prediction for large-scale advanced analytics
Venkataraman, S., Yang, Z., Franklin, M. J. M., Recht, B., & Stoica, I. G. (2016) · Networked Systems Design and Implementation · Cited by 559
Self-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 401–500 of 575
High-Throughput Ingestion for Video Warehouse: Comprehensive Configuration and Effective Exploration
Zhang, B., Li, H., Zhang, D., Li, H., Tan, K., & Chen, G. (2025) · Proceedings of the ACM on Management of Data · doi:10.1145/3725407 · Cited by 2
“…and to construct the Pareto configuration set from the CCS, instead of heuristic algorithms such as traditional hill-climbing used by previous video database systems [4, 13, 19], we employ reinforcement learning (RL) to train an agent to search for configuration for video streams across CCS.”
Harnessing retroreflective transportation infrastructure for intelligent vehicle positioning
Shao, S., & Khreishah, A. A. (2020) · Vehicular Communications · doi:10.1016/j.vehcom.2020.100246 · Cited by 2
“Camera-based vehicle sensing technology [9] is similar to LiDAR, taking into account the difference in terms of sensing capability.”
BackgroundMethodEmbedded Image Analysis System Based on B-ANN
Ionescu, L. M., Mazare, A. G., Visan, D. A., Lita, A. I., Lita, I., & Şerban, G. (2020) · International Spring Seminar on Electronics Technology · doi:10.1109/isse49702.2020.9121051 · Cited by 2
DVQShare: An Analytics System for DNN-based Video Queries
Fu, H., Tang, S., Yu, C., Li, Y., Sun, J., & Liu, Y. (2021) · IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGrid) · doi:10.1109/ccgrid51090.2021.00026 · Cited by 2
Energy-efficient parking analytics system using deep reinforcement learning
Rezaei, Y., Lee, S., & Mossé, D. (2021) · ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation (BuildSys) · doi:10.1145/3486611.3486660 · Cited by 2
A Data Protection Focused Adaptation Engine for Distributed Video Analytics Pipelines
Lachner, C., Laufer, J., Dustdar, S., & Pohl, K. (2022) · IEEE Access · doi:10.1109/access.2022.3185990 · Cited by 2
“In [35], they introduced a controller that dynamically picks the best configurations for existing Neural-Network-based Video Analytics Pipeline.”
Lee, K., & Youn, C. (2022) · International Conference on Information and Communication Technology Convergence (ICTC) · doi:10.1109/ictc55196.2022.9952934 · Cited by 2
“For instance, [10] experimentally showed that if the cluster of the input streams share similar input characteristics, they also show similar profiling pattern.”
BackgroundMethodInfluentialQuerying Zero-streaming CamerasPreprint
Xu, M., Xu, T., Liu, Y., Liu, X., Huang, G., & Lin, F. X. (2019) · arXiv (Cornell University) · Cited by 2
Suprem, A., Lima, R. A., Padilha, B., Ferreira, J. E., & Pu, C. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1912.04423 · Cited by 2
“The approach in [1] uses both YOLO and Mask-RCNN for object detection [1].”
BackgroundMethodInfluentialEKO: Adaptive Sampling of Compressed Video DataPreprint
Bang, J., Chunduri, P., & Arulraj, J. J. P. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2104.01671 · Cited by 2
“Researchers have proposed sampling algorithms that better leverage the temporal redundancy of the video [16, 18, 24].”
BackgroundKong, Z. J., Xu, Q., & Hu, Y. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2507.18748 · Cited by 2
“First, video analytics inference requests often have stringent service level objectives (SLOs), e.g., 200 ms [26, 65].”
EdgeSync: Accelerating Edge-Model Updates for Data Drift through Adaptive Continuous LearningPreprint
Donga, R., Zhao, P., Wang, G., Qi, N., & Lin, J. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.21781 · Cited by 2
“To balance accuracy and processing speed, Jiang et al. [10] designed a controller that dynamically adjusts parameters within the video analysis system.”
Adaptive Acceleration of Inference Services at the Network Edge
Mofjeld, C., Peng, Y., & Qin, H. (2022) · IEEE International Conference on Mobile Cloud Computing, Services, and Engineering (MobileCloud) · doi:10.1109/mobilecloud55333.2022.00009 · Cited by 1
“Others focused on dynamically adapting the quality of the inference [11]–[13].”
BackgroundMethodRoadside Traffic Monitoring Using Video Processing on the Edge
Jaipuria, S., Banerjee, A., & Bhattacharya, A. (2024) · International Conference on Communication Systems & Networks (COMSNETS) · doi:10.1109/comsnets59351.2024.10427468 · Cited by 1
“The time to process the assigned tiles is fixed to L = 16 s for all cases as in prior systems like Chameleon [28], the execution time varies with the number of cameras.”
BackgroundMethodROIAdaptor: Adaptive Task Offloading of ROI-Encoded Videos for Edge Video Analytics
Xu, Z., Guo, X., Huang, Z., Fu, S., & Dong, F. (2023) · International Conference on Mobility, Sensing and Networking (MSN) · doi:10.1109/msn60784.2023.00047 · Cited by 1
“Chameleon [24] considers the resolution of the video, the compression rate of the image to optimize the tradeoff between inference accuracy, task latency, and resource allocation.”
BackgroundDu, C., Tang, S., Song, M. J., Gou, J., Yu, C., Li, Y., Fu, H., Tian, Y., & Yuan, D. (2024) · IEEE International Conference on Cluster Computing (CLUSTER) · doi:10.1109/cluster59578.2024.00042 · Cited by 1
An Analysis of Network Overhead in Distributed TinyML
Hollingsworth, K., Nian, S., Gutierrez, A., & Padmanabhan, A. (2024) · IEEE/ACM Symposium on Edge Computing (SEC) · doi:10.1109/sec62691.2024.00051 · Cited by 1
SDFMD: Spatial Domain Filtering for Accelerating Video Analysis
Sun, H., Liu, W., Liu, M., Gu, Y., Xia, F., Liu, K., Song, Y., Cheng, H., Lei, T., & Zhang, S. (2025) · Lecture notes in electrical engineering · doi:10.1007/978-981-96-5318-8_30 · Cited by 1
EAR-Mapping: Edge-Assisted Real-Time Dense Mapping with Low Bandwidth Requirements
Dai, Y., Qian, B., Yan, Y., Zhao, M., Liu, Y., & Shu, Y. (2025) · International Conference on Mobile Systems, Applications, and Services · doi:10.1145/3711875.3729127 · Cited by 1
Querying large video datasets: a systematic literature review
Kossoski, C., Lopes, H. S., & Simão, J. M. (2025) · Journal of Internet Services and Applications · doi:10.5753/jisa.2025.5495 · Cited by 1
Wang, X., Yan, Y., Zhang, S., Zhu, A., Chen, N., Chen, Y., Qian, Z., Lu, S., & Liang, Y. (2025) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3746027.3755186 · Cited by 1
“Empiri-cally, we observe roughly half of the loss is from result completion, which is inevitable in filtering-based VAPs [1, 19, 24, 26] (accuracy of pseudo frame is depicted in § 4.4).”
InfluentialParachute: Dynamic Resource-Aware Privacy-Preserving Video Analytics on Edge
Xu, W., Yang, S., Li, F., Zhu, L., Zhu, K., Chen, X., & Wang, Y. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3638416 · Cited by 1
“Chameleon [37] introduces a scalable controller that selects neural network configurations based on temporal and spatial correlations across video feeds.”
DACC: Discerning and adaptive offloading for coarse-grained content-aware video analytics
Pan, H., Chen, N., Huang, H., Sun, Y.-E., Wang, X., Xing, Y., Zhang, S., & Wu, J. (2026) · Computer Networks · doi:10.1016/j.comnet.2026.112130 · Cited by 1
Odeh, J. A., & Usman, M. (2026) · African Journal of Management and Business Research · doi:10.62154/ajmbr.2026.022.01018 · Cited by 1
Zhao, K., Wang, Y., Zhou, Z., Zhou, R., Zhang, X., Chen, X., Ran, D., Zhang, F., & Cao, L. (2026) · IEEE Transactions on Services Computing · doi:10.1109/tsc.2026.3688483 · Cited by 1
“By adjusting data transmission parameters (e.g., image resolution or video frame rate for object detection), the amount of transmitted data can be effectively reduced, thus decreasing service response time while potentially impacting DNN inference accuracy [23].”
MERCI: Adaptive Multi-Expert Inference for Dynamic and Large-Vocabulary Vision Perception
Kim, W., Jeong, M., Shin, S. H., Yi, J., & Lee, Y. (2026) · IEEE International Conference on Pervasive Computing and Communications · doi:10.1109/percom67906.2026.11524520 · Cited by 1
“Conventional video analytics systems control various knobs of the input video, e.g. , resolution, frame rate, and quantization parameter [12], with efficient profiling [13] or feedback [14].”
SAIL: Redesigning Collaborative Language Inference with a Single Server-to-Mobile Handoff
Park, G., Han, S., Cho, Y., Park, C., & Lee, K. (2026) · International Conference on Mobile Systems, Applications, and Services · doi:10.1145/3745756.3809224 · Cited by 1
MoVi: Real-Time Large Multimodal Model-Driven Interactive Video Analytics on Mobile Devices
Li, Z., Fan, X., Hu, X., & Zhu, Y. (2026) · IEEE Conference on Computer Communications (INFOCOM) · doi:10.1109/infocom59046.2026.11571484 · Cited by 1
“…and CASVA [3], [15] implement adaptive streaming policies; OAVS, RIVA propose online retraining methods for streaming policies [39], [40], Janus [41] proposes collaborative model splitting solution for vision transformers, and Chameleon [2] optimizes resource usage for large-scale analytics.”
InfluentialZhu, Y., Wen, Z., Wang, T., Yang, Z., Zhang, X., Hong, Z., Qian, B., He, S., Qian, L., & Wang, C. (2026) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3831654 · Cited by 1
Varisetty, T., & Dietrich, D. (2019) · doi:10.1145/3349614.3356021 · Cited by 1
“Examples are provided in the work of [7] and [5], that addresses the issue with network bandwidth for improving the scalability, overhead associated with large video sets, the application of lightweight video analytics using WebSockets for adaptive streaming of 3D interactive video and camera…”
BackgroundApproximate Query Processing on Autonomous Cameras.
Xu, M., Zhang, X., Liu, Y., Liu, X., & Lin, F. X. (2019) · Cited by 1
A Cascade Collaborative Offloading Framework for Video Analytics Based on Online Learning
li, Y., luo, B., Zhang, Y., Sun, Z., & Liu, Y. (2020) · IEEE International Conference on Internet of Things (iThings) and IEEE Green Computing and Communications (GreenCom) and IEEE Cyber, Physical and Social Computing (CPSCom) and IEEE Smart Data (SmartData) · doi:10.1109/ithings-greencom-cpscom-smartdata-cybermatics50389.2020.00138 · Cited by 1
“Furtherly, different from previous works with reinforcement learning[28] and online training parameters[16] or profile optimization[32], [33], we adopt a periodic online learning algorithm to update the weights of the micro-classifier and the threshold of the difference detector adaptively.”
MethodA Low-Cost Video Analytics System with Velocity Based Configuration Adaptation in Edge Computing
Kim, W., & Youn, C. (2021) · Transactions on computational science and computational intelligence · doi:10.1007/978-3-030-70296-0_48 · Cited by 1
The case for admission control of mobile cameras into the live video analytics pipeline
Faticanti, F., Bronzino, F., & Pellegrini, F. D. (2021) · doi:10.1145/3477083.3480151 · Cited by 1
Characterizing the Performance of Deep Learning Inference for Edge Video Analytics
Luo, Z., Wang, X., Zhang, Y., Wang, Z., Hu, M., & Wu, D. (2022) · Communications in computer and information science · doi:10.1007/978-981-16-9709-8_20 · Cited by 1
Low-Latency Speculative Inference on Distributed Multi-Modal Data Streams
Li, T., Huang, J., Risinger, E., & Ganesan, D. (2022) · GetMobile Mobile Computing and Communications · doi:10.1145/3568113.3568121 · Cited by 1
“The results are promising: in both scenarios, our system offers 7-128x latency speedup while achieving the same accuracy against six baselines [5-7].”
BackgroundMultitask Scheduling of Computer Vision Workload on Edge Graphical Processing Units
Bhattacharya, A., Shukla, P., Banerjee, A., Jaipuria, S., Narendra, N. C., & Garg, D. (2023) · International Conference on Communication Systems & Networks (COMSNETS) · doi:10.1109/comsnets56262.2023.10041358 · Cited by 1
“We then evaluate this algorithm with a few reasonable baselines via both synthetic and real data and show that our algorithm performs much better in practice.”
MethodAPT: Adaptive Perceptual quality based camera Tuning using reinforcement learning
Paul, S., Rao, K., Coviello, G., Sankaradas, M., Po, O., Hu, Y., & Chakradhar, S. (2022) · International Conference on Internet of Things: Systems, Management and Security (IOTSMS) · doi:10.1109/iotsms58070.2022.10062226 · Cited by 1
Lightweight Collaborative Perception at the Edge
Gokarn, I., Jayarajah, K., & Misra, A. (2023) · Springer eBooks · doi:10.1007/978-3-031-40787-1_8 · Cited by 1
HeteroPush: Communication-Efficient Video Analytics by Scheduling Heterogeneous Filters
Zhang, Y.-N., Zhang, L., Yuan, M., & Song, Q. (2023) · International Conference on Big Data Computing and Communications (BigCom) · doi:10.1109/bigcom61073.2023.00035 · Cited by 1
“Input filtering [13], [14] and inference reuse [15] are two popular technical routes for efficient video analytics.”
BackgroundMACRO: Incentivizing Multi-Leader Game-Based Pareto-Efficient Crowdsourcing for Video Analytics
Chen, Y., Zhang, S., Zhou, Z., Wang, X., Liang, Y., Chen, N., Yan, Y., Xiao, M., Wu, J., Qian, Z., et al. (2024) · IEEE International Conference on Data Engineering (ICDE) · doi:10.1109/icde60146.2024.00029 · Cited by 1
“To complete the video analytics tasks, platforms have been actively engaging in hiring proper workers, sending video data to them, and strategically selecting the configurations [27] (i.e., frame rates, resolutions and models) to maximize their accuracy-based profits.”
MethodOffloaD: Detection Failure-based Scheduler for Offloading Object Detection
Gómez, C., Patkhullaev, D., & Hernandez, A. C. (2024) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3677321 · Cited by 1
Ye, J., Cernuda, J., Rajesh, N., Bateman, K., Yildiz, O., Peterka, T., Nigmetov, A., Morozov, D., Sun, X., Kougkas, A., et al. (2024) · Proceedings of the International Conference on Parallel Processing · doi:10.1145/3673038.3673070 · Cited by 1
Learning-Based Edge-Assisted UAV Object Detection Against Jamming for Extended Reality
Ye, L., Zhang, J., Chen, H., Lin, Z., Li, J., Lv, Z., & Xiao, L. (2024) · IEEE/CIC International Conference on Communications in China (ICCC) · doi:10.1109/iccc62479.2024.10681679 · Cited by 1
Adaptive Streaming Continuous Learning System for Video Analytics
Li, T., Li, Q., Zhang, M., Yuan, Z., & Jiang, Y. (2024) · International Workshop on Quality of Service · doi:10.1109/iwqos61813.2024.10682886 · Cited by 1
Adapting Complex Event Detection to Perceptual Domain Shifts
Wang, B., Briseno, J. D. G., Han, L., Phillips, H., Craighead, J. D., Purman, B., Kaplan, L., & Srivastava, M. B. (2024) · IEEE Military Communications Conference (MILCOM) · doi:10.1109/milcom61039.2024.10773796 · Cited by 1
SmartTBD: Smart Tracking for Resource-constrained Object Detection
Zhou, S., Hernandez, A. C., Gómez, C., Yin, W., & Björkman, M. (2025) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3703912 · Cited by 1
“Scheduling strategies that aim to optimize the execution of detection and tracking modules by smartly measuring the system and ,modeling its capabilities have also been proposed [4, 12, 15, 25, 32–35].”
Sharma, A., Subasharan, V., Gulati, M., Wanniarachchige, D., & Misra, A. (2025) · ACM Transactions on Internet of Things · doi:10.1145/3736420 · Cited by 1
“Chameleon [20] investigated how cross-camera correlations could be used to amortize the cost of node and application-specific profiling.”
Reinventing Data Stores for Video Analytics.Preprint
Xu, T., Botelho, L. M., & Lin, F. X. (2018) · arXiv (Cornell University) · Cited by 1
MultiScope: Efficient Video Pre-processing for Exploratory Video AnalyticsPreprint
Bastani, F., & Madden, S. R. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2103.14695 · Cited by 1
“We now evaluate MultiScope on 7 diverse video datasets against three baselines: Chameleon [8], BlazeIt [9], and Miris [1].”
BackgroundMethodInfluentialPerceptual Optimizations for Video Capture, Processing, and Storage Systems
Mazumdar, A. (2020) · ResearchWorks at the University of Washington (University of Washington) · Cited by 1
Energy-Efficient Classification at the Wireless Edge with Reliability GuaranteesPreprint
Merluzzi, M., Battiloro, C., Lorenzo, P. D., & Strinati, E. C. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2204.10399 · Cited by 1
“As pointed out in [1], [4], several aspects must be tackled when performing edge inference tasks in the most efficient way, such as model compression techniques to reduce the memory footprint [5], model selection strategies to adaptively select the inference model online [6], or application oriented optimization, for…”
Zhang, H. (2022) · DR-NTU (Nanyang Technological University) · doi:10.32657/10356/160371 · Cited by 1
Gupta, R., Zhao, L., Li, J., Vakhniuk, V., Danilov, C., Eckhardt, J., Bernard, K., & Nahrstedt, K. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2401.15288 · Cited by 1
Exploring DRAM Cache Prefetching for Pooled MemoryPreprint
Tirumalasetty, C., & Annapreddy, N. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2406.14778 · Cited by 1
“Many of these ML techniques from large language models [19], [34], [61], [62] to video processing [29], [32], [63] and others rely on large amounts of data for training and sometimes for retrieval.”
BackgroundCloudEye: A New Paradigm of Video Analysis System for Mobile Visual ScenariosPreprint
Cui, H. Y., Li, Q., Wang, H., & jiang, Y. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2410.18399 · Cited by 1
“Deep learning visual models generally have a large number of parameters[20].”
BackgroundDMVC: Multi-Camera Video Compression Network aimed at Improving Deep Learning AccuracyPreprint
Cui, H. Y., Li, Q., Wang, H., & Jiang, Y. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2410.18400 · Cited by 1
“Humans prioritize visual quality, while machines require precision in deep learning features for enhanced task performance [10, 18].”
BackgroundUrban Sensing Environments: Exploiting the Adaptation Space for Data Protection
Lachner, C. (2022) · reposiTUm (TU Wien) · doi:10.34726/hss.2022.107688 · Cited by 1
Enabling Quality-of-Service in Emerging Workloads
Tariq, A. R. (2024) · CU Scholar (University of Colorado Boulder) · Cited by 1
Sela, G.-E., Agrawal, K. K., Balaji, B., Gonzalez, J., & Stoica, I. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2512.24386 · Cited by 1
“In DMSS, the video is divided into segments (e.g., 2 seconds [46]), and a different model size is used to process each video segment.”
InfluentialSelf-citationData-driven Management Solution for Microservice-based Deep Learning Applications
Houmani, Z. (2021) · theses.fr (ABES) · doi:10.70675/8a393657zad8cz4129za1d0z77b80e884d8b · Cited by 1
Atlas: Optimizing Deployment of Compound AI Workflows on Heterogeneous ClustersPreprint
Gravara, M., Stanisic, A., & Nastic, S. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2609.04513 · Cited by 1
Core utility development for hysia performance optimization
Zhou, S. (2019) · DR-NTU (Nanyang Technological University) · doi:10.32657/10356/82835
“The challenge is if there is any method to preserve or nearly preserve the detection accuracy as well as save resource? An interesting fact is that a more powerful model (such as the Faster-RCNN-ResNet-101 model) performs nearly the same compared to a less powerful model (such as the Faster-RCNN-ResNet-50) on a scene…”
MethodBumbleBee: Application-aware adaptation for container orchestrationPreprint
Lee, H., Noghabi, S. A., Noble, B. D., Furlong, M., & Cox, L. P. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2008.11868
“For example, machine-learning (ML) applications can adapt to bandwidth drops by sending lower-quality data to an inference model, and can adapt to increased network latency by sending data to a faster, lower-quality model [18, 20].”
BackgroundArtificial Intelligence Inference in Edge
Wang, X., Han, Y., Leung, V. C. M., Niyato, D. T., Yan, X., & Chen, X. (2020) · Springer eBooks · doi:10.1007/978-981-15-6186-3_5
Lessons Learned and Open Challenges
Wang, X., Han, Y., Leung, V. C. M., Niyato, D. T., Yan, X., & Chen, X. (2020) · Springer eBooks · doi:10.1007/978-981-15-6186-3_9
Clique: Spatiotemporal Object Re-identification at the City ScalePreprint
Xu, T., Shen, K., Fu, Y. L., Shi, H., & Lin, F. X. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2012.09329 · Cited by 1
“enable resource/data sharing among cameras. There has been many works proposed to optimize video analytics with operator cascades [15,18,29,52], and format tuning to trade accuracy for cost-efficiency [22,25,43,61,63]. Focus [18] saves cost by pre-processing videos with cheap NNs at ingestion. Notably, it clusters obje…”
Creating Robust Deep Neural Networks With Coded Distributed Computing for IoT SystemsPreprint
Hadidi, R., Cao, J., & Kim, H. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2104.04447
Towards quality-centric design and evaluation of big data cyber security analytics systems
Ullah, F. (2020) · Adelaide Research & Scholarship (AR&S) (University of Adelaide)
Learning-based Adaptive Video Streaming
Jiang, X. (2020) · Institutional Repositories DataBase (IRDB)
Real-time Video Analytics at Scale: Design, Deployment and Evaluation
Pakdaman, S. M. (2020) · TSpace (University of Toronto)
Multimedia Edge ComputingPreprint
Wang, Z., Zhu, W., Sun, L., Hu, H., Ma, G., Ma, M., Pang, H., Ye, J., & Li, H. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2105.02409
“proposed to use edge devices to extract global information that is unavailable for a single client, to assist the deployment of deep learning applications [23], so they can reduce the redundant computation of each stream.”
BackgroundResource Management for Data Analytics in Edge Computing Networks
Galanopoulos, A. (2021) · Arrow@dit (Dublin Institute of Technology)
Hidden Markov based Truth Discovery for Multi-Agent Labeling
Jiang, S., & Zhang, L. (2021) · International Conference on Big Data Computing and Communications (BigCom) · doi:10.1109/bigcom53800.2021.00038
LinkStream: A Liquidity Modeling System on Large-Scale Video Stream in Oilfield
Yuan, H., Ma, Q., Hu, Z., Li, X., Wang, X., & Yang, Z. (2021) · IEEE International Conference on Parallel and Distributed Systems (ICPADS) · doi:10.1109/icpads53394.2021.00103
MoniSys: Sensing, Communication, and Video Analysis in UAV Monitoring System
Wang, W. (2022) · eDiss (Georg-August-Universität Göttingen) · doi:10.53846/goediss-9463
Few-Shot Correlation Estimation for Cross-Camera Video Analytics: A Mean-Field Game Approach
Chen, K., Zhu, Y., Kang, Y., & Han, Z. (2022) · IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC) · doi:10.1109/pimrc54779.2022.9977946
“Prior works on video analytics mainly focus on a single stream [5], [6], [7], or optimizing collaborative video analytics from the computing resource management’s perspective [3], [8].”
BackgroundAFR: An Efficient Buffering Algorithm for Cloud Robotic Systems
Wang, Y.-P., Wang, H., Zou, Z., & Manocha, D. (2022) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros47612.2022.9981400
“Chameleon [10] periodically searches for the optimal configuration for a video query, balancing accuracy and computation resources.”
BackgroundToward cooperative multi-agent video streaming perception
杨, 铮., 董, 亮., & 蔡, 新. (2023) · Scientia Sinica Informationis · doi:10.1360/ssi-2021-0179
“其中一个分辨率和采样帧率的组 合被叫做视频配置 [31] , 而不同的视频配置会导致不同的分析精度以及资源消耗 .”
Semantic Fast-Forwarding for Video Training Set Construction
Feng, Z., & Satyanarayanan, M. (2023) · IEEE Workshop on Mobile Computing Systems and Applications (WMCSA) · doi:10.1145/3572864.3580331
“These systems also assume a golden System Search targets considered NoScope [13] Person, bus, car BlazeIt [12] Bus, car, boat Chameleon [11] Targets from the COCO data set [15] Focus [6] Targets from the COCO data set Mainstream [10] Pedestrian, bus, car, train FilterForward [2] Pedestrian,…”
BackgroundVISTA: Fast and Efficient Traffic Surveillance by Tile SamplingPreprint
Chaudhary, S., Taneja, A., Singh, A., Sikdar, S., Maity, M., & Bhattacharya, A. (2023) · doi:10.36227/techrxiv.22083461
FAST: Fast and Accurate Adaptation in Live Video Analytics Using Intermediate Features
Shin, S. H., Yi, J., Jeong, M., & Lee, Y. (2023) · ACM Workshop on Mobile Immersive Computing, Networking, and Systems (ImmerCom) · doi:10.1145/3615452.3617939
Latency-Critical Inference Serving for Deep Learning
Nigade, V. (2023) · doi:10.5463/thesis.308
Qiang, W. (2024) · Journal of Circuits Systems and Computers · doi:10.1142/s0218126624501901
Improving Live Augmented Reality With Neural Configuration Adaptation
Chen, N., Zhang, S., & Lu, S. L. (2023) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3693-0230-9.ch007
DeepSlicing: Collaborative and Adaptive CNN Inference With Low Latency
Zhang, S., Chen, Y., Zhang, S., & Chen, Z. (2023) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3693-0230-9.ch006
A Framework for Tile Processing on Edge Servers for Roadside Traffic Surveillance
Jaipuria, S., Bhattacharya, A., & Banerjee, A. (2023) · International Conference on Algorithms, Computing and Systems (ICACS) · doi:10.1145/3631908.3631927
A Liquidity Analysis System for Large-scale Video Streams in the Oilfield
Ma, Q., Yuan, H., Hu, Z., Wang, X., & Yang, Z. (2024) · ACM Transactions on Sensor Networks · doi:10.1145/3649222
Adaptive Overlap Padding and Resolution Selection for Frame Split-based Edge Video Analytics
Zhu, H., Huang, Z., Guo, X., Liu, M., Chen, B., & Dong, F. (2023) · IEEE International Conference on Parallel and Distributed Systems (ICPADS) · doi:10.1109/icpads60453.2023.00326
“Chameleon [13] dynamically selects the optimal input resolution configuration and periodically analyzes video streams and utilizes video similarity to reduce analysis costs.”
BackgroundMethodColor-based Lightweight Utility-aware Load Shedding for Real-Time Video Analytics at the Edge
Gupta, H., Saurez, E., Röger, H., Bhowmik, S., Ramachandran, U., & Rothermel, K. (2024) · ACM International Conference on Distributed and Event-Based Systems (DEBS) · doi:10.1145/3629104.3666037
Environment Understanding with EdgeSLAM
Xu, J., Yang, Z., Liu, Y., & Cao, H. (2024) · Springer eBooks · doi:10.1007/978-981-97-3573-0_7
Edge Computing in Mobile Visual Systems
Xu, J., Yang, Z., Liu, Y., & Cao, H. (2024) · Springer eBooks · doi:10.1007/978-981-97-3573-0_2
EdgeSLAM 1.0: Architectural Innovations in Mobile Visual SLAM
Xu, J., Yang, Z., Liu, Y., & Cao, H. (2024) · Springer eBooks · doi:10.1007/978-981-97-3573-0_3
Wang, Z., Ji, Y., Zhou, J., Jin, M., & Zhu, M. (2024) · International Conference on Mobile Internet, Cloud Computing and Information Security (MICCIS) · doi:10.1109/miccis63508.2024.00034
“1) Analytics Accuracy: Previous research studies [12, 19] have provided two key insights into characterizing the relationship between configuration and analytics accuracy: ( i ) adjusting frame resolution and sampling frame rates have distinct, independent effects on accuracy; ( ii ) the…”
BackgroundFlow Control Solution to Avoid Bottlenecks in Edge Computing for Video Analytics
Vainio, A., & Tarkoma, S. (2024) · International Conference on Fog and Mobile Edge Computing (FMEC) · doi:10.1109/fmec62297.2024.10710217
UnifiedNN: Efficient Neural Network Training on the CloudPreprint
Taki, S. U., Padmanabhan, A., & Mastorakis, S. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2408.01331
“In video analytics use cases, frameworks, such as Chameleon (Jiang et al. 2018b), profile configurations and adopt the most effective configuration for maximum resource utilization.”
BackgroundMethodScalable and Sustainable Video Analytics on Edge using Sensor Clustering
Chaudhary, S., Bhattacharya, A., Anand, S., & Balasubramanian, A. (2024) · ACM Annual International Conference on Mobile Computing and Networking (MobiCom) · doi:10.1145/3636534.3695902
Vickrey Auction Offloading for Edge-Assisted Video Analytics with Dynamic Gain Prediction
Du, M., Tan, X., Pan, Y., Wang, M., Wang, S., & Zheng, Q. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-2864-3_8
Yang, J. M., Hwang, S., Park, J., & Bahk, S. (2025) · IEEE Access · doi:10.1109/access.2025.3558251
“On-device approach seeks to optimize and best use the limited computational resource [6], [7], [30].”
Langevin Monte Carlo-based Offloading for Edge-Assisted Real-Time Video Analytics
Xu, W., Dai, P., & Zhao, K. (2025) · World Conference on Computing and Communication Technologies (WCCCT) · doi:10.1109/wccct65447.2025.11028004
“Therefore, following established best practices from prior research [5], [6], we utilize outputs from the advanced object detection model deployed on the edge server as an approximation of ground-truth.”
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.