MARS: mmWave-based Assistive Rehabilitation System for Smart Healthcare

Abstract

Rehabilitation is a crucial process for patients suffering from motor disorders. The current practice is performing rehabilitation exercises under clinical expert supervision. New approaches are needed to allow patients to perform prescribed exercises at their homes and alleviate commuting requirements, expert shortages, and healthcare costs. Human joint estimation is a substantial component of these programs since it offers valuable visualization and feedback based on body movements. Camera-based systems have been popular for capturing joint motion. However, they have high-cost, raise serious privacy concerns, and require strict lighting and placement settings. We propose a millimeter-wave (mmWave)-based assistive rehabilitation system (MARS) for motor disorders to address these challenges. MARS provides a low-cost solution with a competitive object localization and detection accuracy. It first maps the 5D time-series point cloud from mmWave to a lower dimension. Then, it uses a convolution neural network (CNN) to estimate the accurate location of human joints. MARS can reconstruct 19 human joints and their skeleton from the point cloud generated by mmWave radar. We evaluate MARS using ten specific rehabilitation movements performed by four human subjects involving all body parts and obtain an average mean absolute error of 5.87 cm for all joint positions. To the best of our knowledge, this is the first rehabilitation movements dataset using mmWave point cloud. MARS is evaluated on the Nvidia Jetson Xavier-NX board. Model inference takes only 64 s and consumes 442 J energy. These results demonstrate the practicality of MARS on low-power edge devices.

Fast and Scalable Human Pose Estimation using mmWave Point Cloud

“…It employs a human shape model to strengthen the ability of deep learning models to predict human shape with fewer points. Finally, another recent study proposes a mmWave-based assistive rehabilitation system (MARS) [2] using human pose estimation. It sorts the mmWave point cloud and performs matrix transformations before feeding them to a CNN model.…”

“…Convolutional neural networks (CNNs) have become the mainstream method to process images and videos due to their ability to effectively extract feature maps from raw data [5]. Likewise, previous mmWave pose estimation studies [2,14,21,22]…”

“…We employ an open-source mm-Wave point cloud dataset (MARS [2]) to evaluate the proposed FUSE framework. The dataset consists of 40,083 labeled frames (defined in Equation 2) collected using TI IWR1443 Boost mmWave radar [19].…”

“…Then, each movement data is individually split into 60% training, 20% validation, and 20% test sets. Baseline ML model: We implement the CNN trained with the MARS dataset [2] as the baseline model to ensure a fair comparison.…”

Residual Channel Attention Network for Brain Glioma Segmentation

“…Early works have approached the problem with classic machine learning approaches, e.g., logistic regression [2], combinations of Bayesian classification and Markov Random Field [3], and hierarchical approaches [4]. In recent years, deep learning methods such as deep convolutional neural networks (DCNNs) have been validated to be more effective in feature recognition on numerous biomedical applications, including disease diagnosis [5], [6], [7], health monitoring [8], [9], [10], [11], biomedical image analysis [12], [13], [14], [15], [16], and Electroencephalography (EEG) analysis [17], [18]. Nevertheless, the problem remains challenging in brain tumor segmentation for various Fig.…”