Change-point Detection Methods for Body-Worn Video

TL;DR

This paper introduces a two-stage framework combining machine learning and statistical algorithms to detect change-points in body-worn video, improving police video analysis efficiency and accuracy.

Contribution

The paper presents a novel two-stage approach integrating scene classification with multiple change-point detection algorithms for BWV analysis.

Findings

Achieved over 90% recall in detecting scene changes.

Nearly 70% precision in change-point detection.

Robustness to scene changes, luminance differences, and occlusions.

Abstract

Body-worn video (BWV) cameras are increasingly utilized by police departments to provide a record of police-public interactions. However, large-scale BWV deployment produces terabytes of data per week, necessitating the development of effective computational methods to identify salient changes in video. In work carried out at the 2016 RIPS program at IPAM, UCLA, we present a novel two-stage framework for video change-point detection. First, we employ state-of-the-art machine learning methods including convolutional neural networks and support vector machines for scene classification. We then develop and compare change-point detection algorithms utilizing mean squared-error minimization, forecasting methods, hidden Markov models, and maximum likelihood estimation to identify noteworthy changes. We test our framework on detection of vehicle exits and entrances in a BWV data set provided by the Los Angeles Police Department and achieve over 90% recall and nearly 70% precision -- demonstrating robustness to rapid scene changes, extreme luminance differences, and frequent camera occlusions.