Probabilistic Activity Recognition for Serious Games with Applications in Medicine

Abstract

Human activity recognition plays an important role especially in medical applications. This paper proposes a formal approach to model such activities, taking into account possible variations in human behavior. Starting from an activity description enriched with event occurrence probabilities, we translate it into a corresponding formal model based on discrete-time Markov chains (DTMCs). We use the PRISM framework and its model checking facilities to express and check interesting temporal logic properties (PCTL) concerning the dynamic evolution of activities. We illustrate our approach on the model of a serious game used by clinicians to monitor Alzheimer patients. We expect that such a modeling approach could provide new indications for interpreting patient performances. This paper addresses only the model definition and its suitability to check behavioral properties of interest. Indeed, this is mandatory before envisioning any clinical study.

Probabilistic model checking for human activity recognition in medical serious games

“…This paper is an updated and extended version of [9]. It largely extends the conference paper with…”

“…Our first contribution is a formal modeling framework where activities are represented by (hierarchical) discrete-time Markov chains whose edges can be decorated with probabilities [9]. Markov chains are deterministic and do not impose to associate a real duration with each action, contrarily to, e.g., timed automata.…”

“…As a second contribution, we have implemented serious game activities as discrete-time Markov chains using the PRISM language [10,9]. We used temporal logic to encode some relevant properties on their dynamical evolution, and we applied model checking techniques [11] to automatically validate the models with respect to these properties as well as to infer the probabilities of some interesting paths.…”

Engaging Serious Games for Energy Efficiency

“…The efficacy of serious games has been demonstrated across many differing areas including, for example, healthcare monitoring [15] evacuation [16], way-finding in an emergency [17], industrial safety [18], and rehabilitation settings [19]. Serious games have been found to be particularly relevant in an safety-critical industries, due to the significant financial and time costs associated with safety drills [20], and low generalisation of safety behaviour to real-world contexts due to limited feedback on outcomes [21] and poor emotional investment [22].…”

Probabilistic Model Checking for Activity Recognition in Medical Serious Games

“…In the line of [6], we expect that our approach to categorize the behavior of patients could provide doctors with significant information to complete their diagnosis. We chose to represent serious game activities as discrete-time Markov chains whose edges can be decorated with probabilities [7]. All the possible actions do not occur with the same probability, since some behaviors are most frequent than others.…”