# Video Analytics with Zero-streaming Cameras

## TL;DR

This paper introduces DIVA, a system for efficient querying of large videos stored locally on low-cost cameras, using novel techniques to optimize resource use and query speed, achieving over 100x real-time performance.

## Contribution

DIVA is the first system enabling efficient querying of large videos on low-cost remote cameras through innovative sparse landmark frames and multi-pass query processing.

## Key findings

- DIVA achieves over 100x real-time video query performance.
- It outperforms alternative designs on diverse queries over 720 hours of video.
- DIVA effectively balances resource use and query accuracy during execution.

## Abstract

Low-cost cameras enable powerful analytics. An unexploited opportunity is that most captured videos remain "cold" without being queried. For efficiency, we advocate for these cameras to be zero streaming: capturing videos to local storage and communicating with the cloud only when analytics is requested. How to query zero-streaming cameras efficiently? Our response is a camera/cloud runtime system called DIVA. It addresses two key challenges: to best use limited camera resource during video capture; to rapidly explore massive videos during query execution. DIVA contributes two unconventional techniques. (1) When capturing videos, a camera builds sparse yet accurate landmark frames, from which it learns reliable knowledge for accelerating future queries. (2) When executing a query, a camera processes frames in multiple passes with increasingly more expensive operators. As such, DIVA presents and keeps refining inexact query results throughout the query's execution. On diverse queries over 15 videos lasting 720 hours in total, DIVA runs at more than 100x video realtime and outperforms competitive alternative designs. To our knowledge, DIVA is the first system for querying large videos stored on low-cost remote cameras.