mlsanityData to 7 Oct 2026

MadEye: Boosting Live Video Analytics Accuracy with Adaptive Camera Configurations

arXiv 2304.02101 ↗First posted 4 Apr 2023cs.DC cs.CV cs.NIcited by 7

Not on the map. Why?

Abstract (arXiv)

Camera orientations (i.e., rotation and zoom) govern the content that a camera captures in a given scene, which in turn heavily influences the accuracy of live video analytics pipelines. However, existing analytics approaches leave this crucial adaptation knob untouched, instead opting to only alter the way that captured images from fixed orientations are encoded, streamed, and analyzed. We present MadEye, a camera-server system that automatically and continually adapts orientations to maximize accuracy for the workload and resource constraints at hand. To realize this using commodity pan-tilt-zoom (PTZ) cameras, MadEye embeds (1) a search algorithm that rapidly explores the massive space of orientations to identify a fruitful subset at each time, and (2) a novel knowledge distillation strategy to efficiently (with only camera resources) select the ones that maximize workload accuracy. Experiments on diverse workloads show that MadEye boosts accuracy by 2.9-25.7% for the same resource usage, or achieves the same accuracy with 2-3.7x lower resource costs.

Key facts. Cited by 7 papers in the map (data through 7 Oct 2026). Builds on 42 of its 117 references that are in the map. First posted on arXiv 4 Apr 2023.

Builds on

42 of 117 references.

References · 42

  1. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  2. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,029
  3. 2015Distilling the Knowledge in a Neural Networkcited by 10,076
  4. 2015Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networkscited by 9,617
  5. 2016Feature Pyramid Networks for Object Detectioncited by 5,021
  6. 2015SSD: Single Shot MultiBox Detectorcited by 3,125
  7. 2018YOLOv3: An Incremental Improvementcited by 2,652
  8. 2013Visualizing and Understanding Convolutional Networkscited by 2,278
  9. 2017Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inferencecited by 1,548
  10. 2014How transferable are features in deep neural networks?cited by 1,216
  11. 2019EfficientDet: Scalable and Efficient Object Detectioncited by 1,018
  12. 2021ByteTrack: Multi-Object Tracking by Associating Every Detection Boxcited by 812
30 more
  1. 2017Network Dissection: Quantifying Interpretability of Deep Visual Representationscited by 703
  2. 2017Learning Efficient Convolutional Networks through Network Slimmingcited by 673
  3. 2020What is the State of Neural Network Pruning?cited by 578
  4. 2017Measuring Catastrophic Forgetting in Neural Networkscited by 359
  5. 2019Towards Fairness in Visual Recognition: Effective Strategies for Bias Mitigationcited by 252
  6. 2020Scaled-YOLOv4: Scaling Cross Stage Partial Networkcited by 211
  7. 2016Trained Ternary Quantizationcited by 203
  8. 2016Speed/accuracy trade-offs for modern convolutional object detectorscited by 198
  9. 2016Clipper: A Low-Latency Online Prediction Serving Systemcited by 191
  10. 2019Fairness in Deep Learning: A Computational Perspectivecited by 123
  11. 2019Imbalance Problems in Object Detection: A Reviewcited by 118
  12. 2020Image Representations Learned With Unsupervised Pre-Training Contain Human-like Biasescited by 107
  13. 2020Towards causal benchmarking of bias in face analysis algorithmscited by 64
  14. 2020Ekya: Continuous Learning of Video Analytics Models on Edge Compute Serverscited by 58
  15. 2017DeepCache: Principled Cache for Mobile Deep Visioncited by 57
  16. 2017Generating High-Quality Crowd Density Maps using Contextual Pyramid CNNscited by 55
  17. 2018BlazeIt: Optimizing Declarative Aggregation and Limit Queries for Neural Network-Based Video Analyticscited by 41
  18. 2019Scaling Video Analytics on Constrained Edge Nodescited by 36
  19. 2022GEMEL: Model Merging for Memory-Efficient, Real-Time Video Analytics at the Edgecited by 32
  20. 2015Learning Complexity-Aware Cascades for Deep Pedestrian Detectioncited by 18
  21. 2018Euphrates: Algorithm-SoC Co-Design for Low-Power Mobile Continuous Visioncited by 10
  22. 2020Accelerating Deep Learning Inference via Freezingcited by 10
  23. 2021Boggart: Towards General-Purpose Acceleration of Retrospective Video Analyticscited by 10
  24. 2021Privid: Practical, Privacy-Preserving Video Analytics Queriescited by 7
  25. 2021AQuA: Analytical Quality Assessment for Optimizing Video Analytics Systemscited by 6
  26. 2020Analyzing Worldwide Social Distancing through Large-Scale Computer Visioncited by 5
  27. 2021Enhancing Video Analytics Accuracy via Real-time Automated Camera Parameter Tuningcited by 4
  28. 2021SkyQuery: An Aerial Drone Video Sensing Platformcited by 3
  29. 2021Cascaded Residual Density Network for Crowd Countingcited by 2
  30. 2022Using Non-Stationary Bandits for Learning in Repeated Cournot Games with Non-Stationary Demandcited by 2

Built on by

Cited by year

Citing papers per year: 2024: 3; 2025: 4.

7 papers.

  1. 2024AxiomVision: Accuracy-Guaranteed Adaptive Visual Model Selection for Perspective-Aware Video Analyticscited by 4
  2. 2025AI Should Sense Better, Not Just Scale Bigger: Adaptive Sensing as a Paradigm Shiftcited by 3
  3. 2024Enabling Cross-Camera Collaboration for Video Analytics on Distributed Smart Camerascited by 1
  4. 2024AI-Powered CPS-Enabled Vulnerable-User-Aware Urban Transportation Digital Twin: Methods and Applicationscited by 1
  5. 2025Legilimens: Performant Video Analytics on the System-on-Chip Edgecited by 0
  6. 2025ECCO: Leveraging Cross-Camera Correlations for Efficient Live Video Continuous Learningcited by 0
  7. 2025RedunCut: Measurement-Driven Sampling and Accuracy Performance Modeling for Low-Cost Live Video Analyticscited by 0