RESUMEN
This work leverages a custom implementation of a deep neural network-based object detection algorithm to detect people and a set of assistive devices relevant to clinical environments. The object detections form the basis for the quantification of different ambulatory activities and related behaviors. Using features extracted from detected people and objects as input to machine learning models, we quantify how a person ambulates and the mode of ambulation being used.Clinical relevance- This system provides the data required for clinicians and hospitalized patients to work together in the creation, monitoring, and adjustment of ambulatory goals.
Asunto(s)
Aprendizaje Profundo , Humanos , Redes Neurales de la Computación , Algoritmos , Aprendizaje Automático , CaminataRESUMEN
Our work identifies subjects based on their height and the distance between their joints. Using a depth sensing camera, we obtained the position of a person's joints in 3D space relative to each other. The distances between adjacent joints and height of a subject's head are used to create a vector of eight features for an individual to use for identification. Using modified KNN, full and partial feature sets were used to identify subjects. Additionally, our classifier can be utilized to assess ambulation (such as walking's velocity and distance) of subject, when identified.