Real-Time Articulated Hand Pose Estimation Using Semi-supervised Transductive Regression Forests

Tang, Dajun; Yu, Tsz-Ho; Kim, Taekyun

doi:10.1109/iccv.2013.400

Cited by 197 publications

(161 citation statements)

References 21 publications

Supporting

Mentioning

161

Contrasting

Order By: Relevance

“…Our evaluation shows that the use of a global estimator in combination with local refinement, improves on state of the art for estimation of hand joints in the dataset of Tang [1].…”

Section: Introductionmentioning

confidence: 98%

“…To promote realism, Xu [3] incorporates the traits of shadowing and missing depth to the training images, indicative of structured light based depth. While Tang [1] explores introducing real data into training using 1200 manually labelled images. Tang acknowledges that "manually labelled realistic data is extremely costly to obtain" and so combines real and synthetic data using semi supervised learning.…”

Section: B Discriminative Modellingmentioning

confidence: 99%

“…In the recent work of Tang [1] and Keskin [2], comparisons have been drawn between the challenge of hand pose and that of body pose estimation. This is because there are similar requirements for real time performance and accuracy in determining the configuration of an articulated object.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Combining discriminative and model based approaches for hand pose estimation

Krejov

Gilbert

Bowden

2015

2015 11th IEEE International Conference and Workshops on Automatic Face and Gesture Recognition (FG)

View full text Add to dashboard Cite

Abstract-In this paper we present an approach to hand pose estimation that combines both discriminative and modelbased methods to overcome the limitations of each technique in isolation. A Randomised Decision Forests (RDF) is used to provide an initial estimate of the regions of the hand. This initial segmentation provides constraints to which a 3D model is fitted using Rigid Body Dynamics. Model fitting is guided using point to surface constraints which bind a kinematic model of the hand to the depth cloud using the segmentation of the discriminative approach. This combines the advantages of both techniques, reducing the training requirements for discriminative classification and simplifying the optimization process involved in model fitting by incorporating physical constraints from the segmentation. Our experiments on two challenging sequences show that this combined method outperforms the current stateof-the-art approach.

show abstract

“…Our evaluation shows that the use of a global estimator in combination with local refinement, improves on state of the art for estimation of hand joints in the dataset of Tang [1].…”

Section: Introductionmentioning

confidence: 98%

Section: B Discriminative Modellingmentioning

confidence: 99%

See 1 more Smart Citation

Combining discriminative and model based approaches for hand pose estimation

Krejov

Gilbert

Bowden

2015

2015 11th IEEE International Conference and Workshops on Automatic Face and Gesture Recognition (FG)

View full text Add to dashboard Cite

show abstract

“…Multiple ways of creating labelled data have been presented in the past, usually on the expense of additional set-up environments and man-work. For instance, labels have been generated via optimization [29], utilizing multiple cameras, integrating special sensors [31] or in a semisupervised way [14,25]. This is also reflected by the limited * This work was performed while at ETH Zürich.…”

Section: Introductionmentioning

confidence: 99%

How to Refine 3D Hand Pose Estimation from Unlabelled Depth Data?

Dibra

Wolf²,

Öztireli

et al. 2017

2017 International Conference on 3D Vision (3DV)

View full text Add to dashboard Cite

Data-driven approaches for hand pose estimation from depth images usually require a substantial amount of labelled training data which is quite hard to obtain. In this work, we show how a simple convolutional neural network, pre-trained only on synthetic depth images generated from a single 3D hand model, can be trained to adapt to unlabelled depth images from a real user's hand. We validate our method on two existing and a new dataset that we capture, both quantitatively and qualitatively, demonstrating that we strongly compare to state-of-the-art methods. Additionally, this method can be seen as an extension to existing methods trained on limited datasets, which helps on boosting their performance on new ones.

show abstract

“…A number of hand pose estimation methods have been proposed which require the use of synthetic hand models to reason how movements of a real hand result in corresponding variations in the input data [2,3,4,5]. Furthermore the existing 3D reconstruction methods are limited by the noise, relatively smaller size of the hand and the requirement to capture multiple images for producing a complete hand model [6].…”

Section: Introductionmentioning

confidence: 99%

Generating a 3D Hand Model from Frontal Color and Range Scans

Asad

Gentet

Basaru

et al. 2015

2015 IEEE International Conference on Image Processing (ICIP)

View full text Add to dashboard Cite

Realistic 3D modeling of human hand anatomy has a number of important applications, including real-time tracking, pose estimation, and human-computer interaction. However the use of RGB-D sensors to accurately capture the full 3D shape of a hand is limited by self-occlusions, relatively smaller size of the hand and the requirement to capture multiple images. In this paper, we propose a method for generating a detailed, realistic hand model from a single frontal range scan and registered color image. In essence, our method converts this 2.5D data into a fully 3D model. The proposed approach extracts joint locations from the color image using a fingertip and interfinger region detector with a Naive Bayes probabilistic model. Direct correspondence between these joint locations in the range scan and a synthetic hand model are used to perform rigid registration, followed by a thin-plate-spline deformation that non-rigidly registers a synthetic model. This reconstructed model maintains similar geometric properties as the range scan, but also includes the back side of the hand. Experimental results demonstrate the promise of the method to produce detailed and realistic 3D hand geometry.

show abstract

Real-Time Articulated Hand Pose Estimation Using Semi-supervised Transductive Regression Forests

Cited by 197 publications

References 21 publications

Combining discriminative and model based approaches for hand pose estimation

Combining discriminative and model based approaches for hand pose estimation

How to Refine 3D Hand Pose Estimation from Unlabelled Depth Data?

Generating a 3D Hand Model from Frontal Color and Range Scans

Contact Info

Product

Resources

About