Approximate Query Service on Autonomous IoT Cameras

TL;DR

Elf is an energy-efficient runtime for autonomous IoT cameras that provides approximate object counts with bounded errors, optimizing actions through a learning-based planner to operate effectively under energy constraints.

Contribution

The paper introduces Elf, a novel runtime that unifies sampling and counting errors, and employs a learning-based planner for real-time decision-making on energy-constrained cameras.

Findings

Elf achieves 11% average error in object counts.

Narrow errors up to 3.4x smaller than baselines.

Operates effectively with over 1,000 hours of video data.

Abstract

Elf is a runtime for an energy-constrained camera to continuously summarize video scenes as approximate object counts. Elf's novelty centers on planning the camera's count actions under energy constraint. (1) Elf explores the rich action space spanned by the number of sample image frames and the choice of per-frame object counters; it unifies errors from both sources into one single bounded error. (2) To decide count actions at run time, Elf employs a learning-based planner, jointly optimizing for past and future videos without delaying result materialization. Tested with more than 1,000 hours of videos and under realistic energy constraints, Elf continuously generates object counts within only 11% of the true counts on average. Alongside the counts, Elf presents narrow errors shown to be bounded and up to 3.4x smaller than competitive baselines. At a higher level, Elf makes a case for advancing the geographic frontier of video analytics.