Video-based Feedback for Assisting Physical Activity

Abstract

Abstract:In this paper, we explore the concept of providing feedback to a user moving in front of a depth camera so that he is able to replicate a specific template action. This can be used as a home based rehabilitation system for stroke survivors, where the objective is for patients to practice and improve their daily life activities. Patients are guided in how to correctly perform an action by following feedback proposals. These proposals are presented in a human interpretable way. In order to align an action that was performed with the template action, we explore two different approaches, namely, Subsequence Dynamic Time Warping and Temporal Commonality Discovery. The first method aims to find the temporal alignment and the second one discovers the interval of the subsequence that shares similar content, after which standard Dynamic Time Warping can be used for the temporal alignment. Then, feedback proposals can be provided in order to correct the user with respect to the template action. Experimental results show that both methods have similar accuracy rate and the computational time is a decisive factor, where Subsequence Dynamic Time Warping achieves faster results.

Key-Skeleton Based Feedback Tool for Assisting Physical Activity

“…Thus, in order to handle speed variations, a temporal alignment technique is usually applied, e.g., Dynamic Time Warping (DTW) [6]. Contrary to [7], to provide feedback for the current skeleton, we do not apply any temporal alignment technique between the movement being performed and the reference movement. Instead, we uniformly sample the reference movementM into a fixed number of skeletons q, denoted as key-skeletons, represented by φ.…”

STARR - Decision SupporT and self-mAnagement system for stRoke survivoRs Vision based Rehabilitation System

“…We developed algorithms for guiding patients to perform correctly an action or movement by presenting feedback proposals in the form of visual information and humaninterpretable messages. Our proposed models use a 3D human body skeleton representation [18][19][20].…”

“…Feedback on Dynamic Human Actions. We explored the concept of providing feedback to a user moving in front of a depth camera so that (s)he is able to replicate a specific template action [19]. Note that, we denote a template action as a dynamic template action (i.e series of poses) to differ from the previous section, where the static template (i.e a single pose) was considered.…”

“…The interval corresponding to the temporal alignment is represented by the red rectangle. Feedback proposals are shown for the same instance for both methods in order to correct the position with respect to the template action.Figure adapted from[19].…”

Two-Stage RGB-Based Action Detection Using Augmented 3D Poses

“…Nevertheless, due to the use of effective RGB-based human motion descriptors in a large spatio-temporal volume, these approaches are usually demanding in terms of computational time. As a result, they can be barely adapted to real-world applications such as security and video surveillance [1], healthcare [2,3], human-computer interaction [24], etc. Furthermore, they perform poorly in the presence of radial motion (defined as the perpendicular motion to the image plane), since this information is not encoded in 2D descriptors.…”