BodyPressure -- Inferring Body Pose and Contact Pressure from a Depth Image

Henry M. Clever,Patrick Grady,Greg Turk,Charles C. Kemp

DOI: https://doi.org/10.48550/arXiv.2105.09936

2021-05-21

Abstract:Contact pressure between the human body and its surroundings has important implications. For example, it plays a role in comfort, safety, posture, and health. We present a method that infers contact pressure between a human body and a mattress from a depth image. Specifically, we focus on using a depth image from a downward facing camera to infer pressure on a body at rest in bed occluded by bedding, which is directly applicable to the prevention of pressure injuries in healthcare. Our approach involves augmenting a real dataset with synthetic data generated via a soft-body physics simulation of a human body, a mattress, a pressure sensing mat, and a blanket. We introduce a novel deep network that we trained on an augmented dataset and evaluated with real data. The network contains an embedded human body mesh model and uses a white-box model of depth and pressure image generation. Our network successfully infers body pose, outperforming prior work. It also infers contact pressure across a 3D mesh model of the human body, which is a novel capability, and does so in the presence of occlusion from blankets.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to address the problem of inferring human posture and contact pressure from a single image using a low-cost depth camera. Specifically: 1. **Background and Challenges**: - Pressure injuries are common among bedridden patients, but current technologies for detecting pressure injuries (such as pressure mats) are expensive and not widely available. - Using low-cost sensors (such as cameras) to perform this task can significantly reduce the incidence of pressure injuries. - However, perceiving pressure through camera images faces many challenges, including the human contact interface being occluded by the body itself and being covered by blankets, making it difficult to locate. 2. **Proposed Method**: - A method named **BodyPressure** is proposed, which can accurately infer human posture and contact pressure from images captured by a single low-cost downward-facing depth camera. - This method maps pressure onto a human model, allowing the identification of high-pressure areas under the body, thereby helping to prevent pressure injuries. 3. **Technical Implementation**: - Utilizes rapid physical simulation to generate a large amount of synthetic data (BodyPressureSD), including depth images, human postures, and pressure mat data. - Trains a deep neural network (BodyPressureWnet) that takes depth images as input and outputs human posture and contact pressure. - The network includes an embedded human mesh model and uses a white-box model to generate depth and pressure images. 4. **Summary of Contributions**: - Proposes a method to infer human posture and contact pressure from depth images. - Creates a synthetic dataset (BodyPressureSD) containing 97,495 samples. - Creates a 3D human model dataset (SLP-3Dfits) containing 4,545 samples, used to initialize the simulator and evaluate the depth model.

BodyPressure -- Inferring Body Pose and Contact Pressure from a Depth Image

BodyPressure - Inferring Body Pose and Contact Pressure from a Depth Image

Bodies at Rest: 3D Human Pose and Shape Estimation from a Pressure Image using Synthetic Data

BodyMAP -- Jointly Predicting Body Mesh and 3D Applied Pressure Map for People in Bed

3D Human Pose Estimation on a Configurable Bed from a Pressure Image

MassNet: A Deep Learning Approach for Body Weight Extraction from A Single Pressure Image

Estimating Pose from Pressure Data for Smart Beds with Deep Image-based Pose Estimators

Pressure Eye: In-bed Contact Pressure Estimation via Contact-less Imaging

A novel in-bed body posture monitoring for decubitus ulcer prevention using body pressure distribution mapping

In-Bed Posture Classification Using Deep Neural Network

PressureVision: Estimating Hand Pressure from a Single RGB Image

A Privacy-Preserving and Unobtrusive Sitting Posture Recognition System Via Pressure Array Sensor and Infrared Array Sensor for Office Workers

A Method to Recognize Sleeping Position Using an CNN Model Based on Human Body Pressure Image

From Kinematics To Dynamics: Estimating Center of Pressure and Base of Support from Video Frames of Human Motion

A Novel Body Posture Recognition System on Bed

Smart Multimodal In-Bed Pose Estimation Framework Incorporating Generative Adversarial Neural Network

Privacy-Preserving in-bed pose monitoring: A fusion and reconstruction study

SATPose: Improving Monocular 3D Pose Estimation with Spatial-aware Ground Tactility

Deep Residual Networks for Sleep Posture Recognition with Unobtrusive Miniature Scale Smart Mat System

SaccpaNet: A Separable Atrous Convolution-based Cascade Pyramid Attention Network to Estimate Body Landmarks Using Cross-modal Knowledge Transfer for Under-blanket Sleep Posture Classification

Identity and Posture Recognition in Smart Beds with Deep Multitask Learning