Real-Time Human Pose Recognition in Parts from Single Depth Images

Shotton, Jamie; Fitzgibbon, Andrew; Cook, Mat; Sharp, Toby; Finocchio, Mark; Moore, Richard B.; Kipman, Alex; Blake, A.

doi:10.1007/978-3-642-28661-2_5

Cited by 378 publications

(517 citation statements)

References 37 publications

Supporting

Mentioning

515

Contrasting

Order By: Relevance

“…However, the occlusion handling process improves Fig. 15 Upper row: the qualitative comparison of the estimated BP joints of ours (in blue), [33] (in red), and [1] (in green) using the dataset [37]. Lower row: The quantitative comparison of ours and the other methods [1,33] by using the dataset [37] BP Joint esƟmaƟon error comprison using HON4D dataset [37] Ours [33] [1]…”

Section: Comparisonsmentioning

confidence: 99%

“…The depth video camera promotes the development of motion-capturing technology [2,[9][10][11][12][13]22]. Visionbased posture estimation can be categorized as model-based method [2,4,9,[20][21][22], example-based method [5,[11][12][13][14][15], and label-based method [1,19]. The first method uses the prediction-estimation technique to track the limbs of human objects and capture human pose parameters based on the pre-stored human model.…”

Section: Introductionmentioning

confidence: 99%

“…The centroid of each segmented BP region is a skeleton joint. Shotton et al [1] propose a method to localize the body joints from a single depth image. To avoid over-fitting problem, they train a random forest (RF) classifier [16][17][18]24] and adopt GPU acceleration [23] for system implementation.…”

Section: Introductionmentioning

confidence: 99%

“…Different from [1,2], our method can estimate the joint locations of human upper body skeleton to capture the human articulated motion parameters with self-occlusion. The human body motion is categorized into four different action types.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Human body motion parameters capturing using kinect

Hsu

Huang

Kao

et al. 2015

Machine Vision and Applications

View full text Add to dashboard Cite

This paper introduces a new real-time human motion parameters capturing method using Kinect. It consists of five modules. First, the hybrid action type classifier categories human body motion into four different action types. Second, for each action type, there is a body part (BP) classifier which segments the human silhouette into 16 BP regions of which the centroids become the BP joints. These BP joints are linked to represent the human body skeleton. Third, an action type validation process verifies the identified action type. Fourth, the partial occlusion recovery method relocates the occluded BP joints. Fifth, the offset compensation process fine tunes the positions of BP joints and then validates the compensation results. The major contributions of this paper are hybrid action type classification and correction, offset compensation, and partial occlusion recovery. The experimental results show that this method can estimate human upper limb motion parameters in real time accurately and effectively.

show abstract

Section: Comparisonsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Human body motion parameters capturing using kinect

Hsu

Huang

Kao

et al. 2015

Machine Vision and Applications

View full text Add to dashboard Cite

show abstract

“…This system also provides a human pose detection algorithm by applying a classifier over the depth images as provided by the sensor [24].…”

Section: Current Mocap State Of the Artmentioning

confidence: 99%

Garment-based motion capture (GaMoCap): high-density capture of human shape in motion

Biasi

Setti

Bue

et al. 2015

Machine Vision and Applications

View full text Add to dashboard Cite

This paper presents a new motion capture (MoCap) system, the garment-based motion capture system -GaMoCap. The key feature is the use of an easily wearable garment printed with colour-coded pattern and a generic multicamera setup with standard video cameras. The coded pattern allows a high-density distribution of markers per unit of surface (about 40 markers per 100 cm 2 ), avoiding markers-swap errors. The high density of markers reconstructed makes possible a simultaneous reconstruction of shape and motion, which gives several concurrent advantages with respect to the state of the art and providing performances comparable with previous marker-based systems. In particular, we provide effective solutions to counter the soft-tissue artefact which is a common problem for garment-based techniques. This effect is reduced using Point Cluster Technique to filter out the points strongly affected by non-rigid motion. Uncertainty of motion estimation has been experimentally quantified by comparing with a state-of-the-art commercial system and numerically predicted by means of a Monte Carlo Method procedure. The experimental evaluation was performed on three different articulated motions: shoulder, knee and hip flexion-extension. The results shows that for the three The work was supported by the motion angles estimated with GaMoCap, the system provides comparable accuracies against a commercial VICON system.

show abstract

Enhanced flyby science with onboard computer vision: Tracking and surface feature detection at small bodies

Fuchs

Thompson

Bue

et al. 2015

Earth and Space Science

View full text Add to dashboard Cite

Spacecraft autonomy is crucial to increase the science return of optical remote sensing observations at distant primitive bodies. To date, most small bodies exploration has involved short timescale flybys that execute prescripted data collection sequences. Light time delay means that the spacecraft must operate completely autonomously without direct control from the ground, but in most cases the physical properties and morphologies of prospective targets are unknown before the flyby. Surface features of interest are highly localized, and successful observations must account for geometry and illumination constraints. Under these circumstances onboard computer vision can improve science yield by responding immediately to collected imagery. It can reacquire bad data or identify features of opportunity for additional targeted measurements. We present a comprehensive framework for onboard computer vision for flyby missions at small bodies. We introduce novel algorithms for target tracking, target segmentation, surface feature detection, and anomaly detection. The performance and generalization power are evaluated in detail using expert annotations on data sets from previous encounters with primitive bodies.

show abstract

Real-Time Human Pose Recognition in Parts from Single Depth Images

Cited by 378 publications

References 37 publications

Human body motion parameters capturing using kinect

Human body motion parameters capturing using kinect

Garment-based motion capture (GaMoCap): high-density capture of human shape in motion

Enhanced flyby science with onboard computer vision: Tracking and surface feature detection at small bodies

Contact Info

Product

Resources

About