ECCO: Leveraging Cross-Camera Correlations for Efficient Live Video Continuous Learning
This addresses scalability issues in video analytics for real-time applications, offering a resource-efficient solution for continuous learning in multi-camera systems.
The paper tackles the high compute and communication costs of retraining separate DNN models for each camera in live video analytics by proposing ECCO, a framework that leverages cross-camera correlations in data drift to retrain shared models, resulting in a 6.7%-18.1% improvement in retraining accuracy with the same resources or support for 3.3 times more concurrent cameras at the same accuracy.
Recent advances in video analytics address real-time data drift by continuously retraining specialized, lightweight DNN models for individual cameras. However, the current practice of retraining a separate model for each camera suffers from high compute and communication costs, making it unscalable. We present ECCO, a new video analytics framework designed for resource-efficient continuous learning. The key insight is that the data drift, which necessitates model retraining, often shows temporal and spatial correlations across nearby cameras. By identifying cameras that experience similar drift and retraining a shared model for them, ECCO can substantially reduce the associated compute and communication costs. Specifically, ECCO introduces: (i) a lightweight grouping algorithm that dynamically forms and updates camera groups; (ii) a GPU allocator that dynamically assigns GPU resources across different groups to improve retraining accuracy and ensure fairness; and (iii) a transmission controller at each camera that configures frame sampling and coordinates bandwidth sharing with other cameras based on its assigned GPU resources. We conducted extensive evaluations on three distinctive datasets for two vision tasks. Compared to leading baselines, ECCO improves retraining accuracy by 6.7%-18.1% using the same compute and communication resources, or supports 3.3 times more concurrent cameras at the same accuracy.