Traditional approaches are sufficient Belinostat fda neither in terms of continuity nor reliability. For example, sporadic doctor’s visits limit the monitoring continuity and subject self-annotations constraint the reports reliability [6]. The ideal case would be having an expert who continuously monitors the subject without requiring direct human intervention. Activity recognition enables the collection of user information via strategically-deployed autonomous systems. Moreover, collected data support decision making. Thus, activity recognition is a potential solution for patient continuous monitoring. In addition, retrieved information might also allow physicians and researchers to gain insight about the diseases’ origin and the effect of the provided treatments.
Activity recognition is a Inhibitors,Modulators,Libraries highly complex problem due to the several different factors that must be considered. Some of them are directly related to the subjects such as the age, gender, weight or height, introducing substantial Inhibitors,Modulators,Libraries differences among the individuals’ execution styles. Moreover these conducts may be potentially disturbed when the subject suffers from physical or cognitive disabilities and/or diseases. Ambient and context-related factors also complicate the recognition task (e.g., we expect different motion styles when a barefoot subject walks on the grasp than when walking with ski boots). The system designer has to deal with the major part of these issues, however there exist other kind of problems particularly related to the sensor modalities used.The use of cameras has been comprehensively studied in activity recognition (an extensive review is presented in [7]).
However, there exist some drawbacks which may determine them as an inconvenient solution. Beyond important issues directly related to the end-user such as privacy, or particular technological challenges as occlusions, light Inhibitors,Modulators,Libraries absence or Inhibitors,Modulators,Libraries noise, one of the main problems for the use of cameras is the availability of such systems in real scenarios. Daily living activities do not solely circumscribe to households or buildings but also to outdoor environments where it might be complicated to have such sensor modalities. For example, activities such as hiking, cycling or Entinostat skiing might be difficult to monitor with cameras in a realistic context. Since the aim is to perform a continuous monitoring of the individual’s behavior, the system should be always available to monitor the subject under any circumstances.
The problems with the use of cameras are solved through the use of sensors which are permanently on the subject 17-AAG purchase (on-body or wearable sensors). Inertial sensors are the most used modality to support the measurement of the body kinematics. One of the main problems on their use is about the obtrusiveness and discomfort experienced by the subject when wearing them.