Multi-View Hand-Hygiene Recognition for Food Safety

Abstract

A majority of foodborne illnesses result from inappropriate food handling practices. One proven practice to reduce pathogens is to perform effective hand-hygiene before all stages of food handling. In this paper, we design a multi-camera system that uses video analytics to recognize hand-hygiene actions, with the goal of improving hand-hygiene effectiveness. Our proposed two-stage system processes untrimmed video from both egocentric and third-person cameras. In the first stage, a low-cost coarse classifier efficiently localizes the hand-hygiene period; in the second stage, more complex refinement classifiers recognize seven specific actions within the hand-hygiene period. We demonstrate that our two-stage system has significantly lower computational requirements without a loss of recognition accuracy. Specifically, the computationally complex refinement classifiers process less than 68% of the untrimmed videos, and we anticipate further computational gains in videos that contain a larger fraction of non-hygiene actions. Our results demonstrate that a carefully designed video action recognition system can play an important role in improving hand hygiene for food safety.

Designing a Computer-Vision Application: A Case Study for Hand-Hygiene Assessment in an Open-Room Environment

“…This work extends our previous work [3,7] in three main ways. First, we extend hand-hygiene action recognition from a single restroom-sink scenario into food-handling laboratory scenarios with a variety of camera views and background layouts.…”

“…In this paper, as in our previous work [ 3 , 7 ], we consider a different application of hand-hygiene assessment: reducing pathogens in food-handling facilities. It is estimated that each year over 9 million people suffer from a food-bourne illness in the United States alone [ 16 ].…”

“…Our previous work [ 3 ] presented a two-stage multi-camera system for hand-hygiene quality assessment. In it, we focused on hand-hygiene assessment in a constrained restroom environment using both stationary and body-mounted cameras.…”

“…The system we describe in this paper is motivated by several design goals. First, similar to [ 3 , 7 ], we need to detect when, and possibly for how long, the basic hand-hygiene actions occur. Second, the system must perform well in an open-room environment.…”

“…Actions considered, including their categorizations for this study. Note some action categorizations were different in[3,7].…”