Home 3D body scans from noisy image and range data

Weiss, Alexander K. H.; Hirshberg, David A.; Black, Michael J.

doi:10.1109/iccv.2011.6126465

Cited by 266 publications

(193 citation statements)

References 15 publications

Supporting

Mentioning

192

Contrasting

Order By: Relevance

“…Note that 90% of the time in our method is used for computing closest points. Previous work on human body reconstruction [7] can only capture nearly naked human bodies and spends nearly one hour of computation time, and prior work on articulated registration [3] computes the registration frame by frame in K minimization steps, taking nearly two hours to compute.…”

Section: Resultsmentioning

confidence: 99%

“…However, they do not use the color information, and fundamental proplem is that they can not handle non-rigid movement, so the reconstructed results in the arm and leg parts are not of high quality. The work by Weiss et al [7] estimates the body shape by fitting the parameters of a SCAPE model [8] to depth data and image silhouettes from a single Kinect. In contrast, our work does not require a prior shape model and relies mainly on registration.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

KinectAvatar: Fully Automatic Body Capture Using a Single Kinect

Cui¹,

Chang²,

Noll³

et al. 2013

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. We present a novel scanning system for capturing a full 3D human body model using just a single depth camera and no auxiliary equipment. We claim that data captured from a single Kinect is sufficient to produce a good quality full 3D human model. In this setting, the challenges we face are the sensor's low resolution with random noise and the subject's non-rigid movement when capturing the data. To overcome these challenges, we develop an improved superresolution algorithm that takes color constraints into account. We then align the super-resolved scans using a combination of automatic rigid and non-rigid registration. As the system is of low price and obtains impressive results in several minutes, full 3D human body scanning technology can now become more accessible to everyday users at home.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

KinectAvatar: Fully Automatic Body Capture Using a Single Kinect

Cui¹,

Chang²,

Noll³

et al. 2013

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…They formulate the problem as a forward graphics synthesis problem and then differentiate it, paying special attention to obtaining derivatives at object boundaries; we adopt a similar approach. Weiss et al [36] estimate both human pose and shape using range data from Kinect and an edge term corresponding to the boundary of the human body. They formulate a differentiable silhouette edge term and mention that it is sometimes not differentiable, but that this occurs at only finitely many points, which can be ignored.…”

Section: Related Workmentioning

confidence: 99%

OpenDR: An Approximate Differentiable Renderer

Loper

Black

2014

Lecture Notes in Computer Science

Self Cite

346

218

View full text Add to dashboard Cite

Abstract. Inverse graphics attempts to take sensor data and infer 3D geometry, illumination, materials, and motions such that a graphics renderer could realistically reproduce the observed scene. Renderers, however, are designed to solve the forward process of image synthesis. To go in the other direction, we propose an approximate differentiable renderer (DR) that explicitly models the relationship between changes in model parameters and image observations. We describe a publicly available OpenDR framework that makes it easy to express a forward graphics model and then automatically obtain derivatives with respect to the model parameters and to optimize over them. Built on a new autodifferentiation package and OpenGL, OpenDR provides a local optimization method that can be incorporated into probabilistic programming frameworks. We demonstrate the power and simplicity of programming with OpenDR by using it to solve the problem of estimating human body shape from Kinect depth and RGB data.

show abstract

“…Although offering high point accuracy [22] without associated health risks [23], their conventionally cited accuracy may be reduced when scanning living humans due to the possibility of involuntary movement over the lengthy scanning duration (˜30 minutes). Full body scanners based on laser [24], and structured light [25] offer shorter scan times (<30 seconds), but are prohibitively expensive for the majority of sports and healthcare research laboratories [26].…”

Section: Introductionmentioning

confidence: 99%