A Markovian-based Approach for Daily Living Activities Recognition

TL;DR

This paper introduces a hierarchical hidden Markov model and a new activity language to improve recognition of complex daily activities and abnormal behaviors in healthcare settings.

Contribution

It presents a novel hierarchical hidden Markov model and a specialized activity language for better indoor activity recognition and abnormality detection.

Findings

Effective recognition of complex indoor activities

Improved detection of abnormal activities

Enhanced modeling of human behavior in domestic spaces

Abstract

Recognizing the activities of daily living plays an important role in healthcare. It is necessary to use an adapted model to simulate the human behavior in a domestic space to monitor the patient harmonically and to intervene in the necessary time. In this paper, we tackle this problem using the hierarchical hidden Markov model for representing and recognizing complex indoor activities. We propose a new grammar, called "Home By Room Activities Language", to facilitate the complexity of human scenarios and consider the abnormal activities.