Learning Frame-Level Classifiers for Video-Based Real-Time Assessment of Stroke Rehabilitation Exercises from Weakly Annotated Datasets

Autonomous rehabilitation support solutions, such as virtual coaches, should provide real-time feedback to improve motor function and maintain patient engagement. However, fully annotated dataset collection for real-time exercise assessment is time-consuming and costly, posing a barrier to evaluating proposed methods. In this work, we present a novel framework that learns a frame-level classifier using weakly annotated videos for real-time assessment of compensatory motions in stroke rehabilitation exercises by generating pseudo-labels at a frame level. We consider three approaches: 1) a baseline approach that uses a source dataset to train a frame-level classifier, 2) a transfer learning approach that uses target dataset video-level labels and parameters learned from a source dataset with frame-level labels, and 3) a semi-supervised approach that leverages a target dataset video-level labels and a small set of frame-level labels. We intend to generalize to a weakly labeled target dataset with new exercises and patients. To validate the approach, we use two datasets annotated on compensatory motions: TULE, an existing video and frame-level labeled dataset of 15 post-stroke patients and three exercises, and SERE, a new dataset of 20 post-stroke patients and five exercises, created by the authors, with video-level labels and a small amount of frame-level labels. We show that a frame-level classifier trained on TULE does not generalize well on SERE ( f_1 = 72.87% ), but our semi-supervised and transfer learning approaches achieve, respectively, f_1 = 78.93% and f_1 = 80.47% . Generating pseudo-labels leads to better frame-level classification results for the target dataset than training a classifier with the source dataset (baseline). Thus, the proposed approach can simplify the customization of virtual coaches to new patients and exercises with low data annotation efforts.