Kinect unbiased

Martínez, Manuel Martínez; Stiefelhagen, Rainer

doi:10.1109/icip.2014.7026171

Cited by 5 publications

(3 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Based on the results of the research analysis, we focused on the process of sit-to-stand transitions of elderly users; this experiment used the Kinect 2.0 sensor as the experimental device to capture the user’s behavior. The device could be used to capture 3D motion data, and the natural interactions between the machine and the user could be achieved [ 25 , 26 ]. The main functional components of the Kinect sensor included the RGB color camera, the infrared transmitter, and the receiver [ 27 , 28 ].…”

Section: Methodsmentioning

confidence: 99%

Recognition and Analysis of an Age-Friendly Intelligent Sofa Design Based on Skeletal Key-Points

Zhou

Huang

Luo

et al. 2022

IJERPH

View full text Add to dashboard Cite

The aging population has a higher level of consumption willingness, higher quality of life demands, and more diversified spiritual pursuits. In recent years, age-friendly industries have entered a ’blue ocean of development’; intelligent recreation and age-friendly industries have become new growth points in the double-cycle situation; however, the current generalized design of the market is not enough to meet the needs of its elderly users. Through an extensive research and demand analysis focused on the highly-frequent use of sofas by elderly individuals, an AHP analysis showed that smart sofa design guidelines are among the most important functional indicators; user ’pain points’ focused on low seat surfaces caused by the difficulty in getting up. To further refine this action behavior, a Kinect-based experimental device was used to capture the behavior of the elderly user during the sit-to-stand transition. The experimental data were collected from 25 key skeletal points in the human body and further investigated by converting the skeletal points into 12 key joint angles to refine the joint transition threshold risk during the sit-to-stand transition for the elderly user. The test results show that the most important joint angle affecting the sit-to-stand transition process is the trunk thigh angle was θ5-2, with an important value of 0.122. The two-dimensional body data of the elderly user was mapped to the joint angles under the three-dimensional activity threshold to build a comfort model of the elderly user’s sofa, providing a theoretical basis for the design parameters of the aging sofa. In response to the research results, an intelligent age-friendly sofa with three forms was designed and prototyped independently, meeting the universal size of elderly users in terms of dimensional parameters, and iterating leisure and assisted standing transformation in terms of function (to reduce the risk levels of the sitting and standing transformations of elderly users).

show abstract

Section: Methodsmentioning

confidence: 99%

Recognition and Analysis of an Age-Friendly Intelligent Sofa Design Based on Skeletal Key-Points

Zhou

Huang

Luo

et al. 2022

IJERPH

View full text Add to dashboard Cite

show abstract

“…The Kinect system is capable of capturing visual RGB-depth (RGB-D) information and generate real-time depth maps containing discrete range measurements of the physical scene, which can be later reprojected as a set of discrete 3D points. Such depth map generation method has been analyzed by several research groups to determine and tackle its limitations (Martinez and Stiefelhagen 2014). In spite of such limitations, Kinect has been used in a wide spectrum of applications that require the automatic extraction of anthropometric or soft biometric measures, e.g.…”

Section: Related Workmentioning

confidence: 99%

Automatic Sleep System Recommendation by Multi-modal RBG-Depth-Pressure Anthropometric Analysis

Palmero

Esquirol²,

Bayo³

et al. 2016

Int J Comput Vis

View full text Add to dashboard Cite

This paper presents a novel system for automatic sleep system recommendation using RGB, depth and pressure information. It consists of a validated clinical knowledge-based model that, along with a set of prescription variables extracted automatically, obtains a personalized bed design recommendation. The automatic process starts by performing multi-part human body RGB-D segmentation combining GrabCut, 3D Shape Context descriptor and Thin Plate Splines, to then extract a set of anthropometric landmark points by applying orthogonal plates to the segmented human body. The extracted variables are introduced to the computerized clinical model to calculate body circumferences, weight, morphotype and Body Mass Index categorization. Furthermore, pressure image analysis is performed to extract pressure values and at-risk points, which are also introduced to the model to eventually obtain the final prescription of mattress, topper, and pillow. We validate the complete system in a set of 200 subjects, showing accurate category classification and high correlation results with respect to manual measures.

show abstract

“…However, the difficulty is that these points are indistinguishable. Martinez et al [53] proposed a more accurate algorithm for locating points. The algorithm first uses block matching to roughly estimate the location of candidate points, and then applies reverse lookup, clustering and back projection methods to accurately determine the location of the point.…”

Section: Dmcnnmentioning

confidence: 99%

Depth measurement based on a convolutional neural network and structured light

Jia

Liu

Yuan

et al. 2021

Meas. Sci. Technol.

View full text Add to dashboard Cite

Depth measurement methods based on structured light are popular due to their advantages of low cost, good portability and ease of implementation. Depth information for the object is obtained based on the geometric relationship of the imaging system and triangulation theory, which usually requires local stereo matching operations. However, this is computationally intensive, resulting in reduced depth accuracy and worse depth maps. To address these problems, this paper proposes a novel depth measurement method based on a convolutional neural network (DMCNN), which is cast as a pixel-wise classification–regression task without matching. Firstly, the DMCNN network is designed as an encoder–decoder structure. A feature pyramid is adopted in the encoder to extract multi-scale fusion features, and parallel classification and regression branches are constructed at the end of the decoder to achieve depth prediction from coarse to fine. Secondly, we use a four-step phase shift algorithm to generate ground truth depth maps and build a dataset containing a large number of speckle distortion images and their corresponding depth maps to train our network. The algorithm runs on an RTX 2080Ti graphics processing unit (GPU) using 20 000 training images. Experimental results show that our method can achieve higher accuracy than alternative depth measurement methods.

show abstract

Kinect unbiased

Cited by 5 publications

References 5 publications

Recognition and Analysis of an Age-Friendly Intelligent Sofa Design Based on Skeletal Key-Points

Recognition and Analysis of an Age-Friendly Intelligent Sofa Design Based on Skeletal Key-Points

Automatic Sleep System Recommendation by Multi-modal RBG-Depth-Pressure Anthropometric Analysis

Depth measurement based on a convolutional neural network and structured light

Contact Info

Product

Resources

About