Multi-view Body Part Recognition with Random Forests

Kazemi, Vahid; Burénius, Magnus; Azizpour, Hossein; Sullivan, Josephine

doi:10.5244/c.27.48

Cited by 66 publications

(102 citation statements)

References 19 publications

Supporting

Mentioning

101

Contrasting

Order By: Relevance

“…pictorial structures) helps to improve the final result. The results are summarized in Table 1 For most of the body parts, we achieve similar results with the classification forest of [19]. In our formulation, we do not rely on a body prior model for smoothing the results.…”

Section: Football Datasetsupporting

confidence: 61%

“…We perform all the experiments only on the training images of the Image Parse [3] dataset to avoid parameter over-fitting. Then, we compare our method with an approach which relies on body part classification forests on the KTH Football dataset [19]. In order to show the power of our model in comparison to part-based methods, we evaluate on the Image Parse dataset.…”

Section: Methodsmentioning

confidence: 99%

“…In this experiment, we compare our method with the part-based method which relies on classification forests [19]. In this work the forest classifies each pixel in the image as a specific body joint.…”

Section: Football Datasetmentioning

confidence: 99%

“…One idea for improving the model is by using better appearance models [16][17][18]. This has also been done by using Random Forests for body part classification [19] or regression [20]. Shape-based body parts generally achieved better performance [21].…”

Section: Related Workmentioning

confidence: 99%

See 3 more Smart Citations

Holistic Human Pose Estimation with Regression Forests

Belagiannis

Amann

Navab

et al. 2014

Articulated Motion and Deformable Objects

View full text Add to dashboard Cite

Abstract. In this work, we address the problem of human pose estimation in still images by proposing a holistic model for learning the appearance of the human body from image patches. These patches, which are randomly chosen, are used for extracting features and training a regression forest. During training, a mapping between image features and human poses, defined by joint offsets, is learned; while during prediction, the body joints are estimated with an efficient mode-seeking algorithm. In comparison to other holistic approaches, we can recover body poses from occlusion or noisy data. We demonstrate the power of our method in two publicly available datasets and propose a third one. Finally, we achieve state-of-the-art results in comparison to other approaches.

show abstract

Section: Football Datasetsupporting

confidence: 61%

Section: Methodsmentioning

confidence: 99%

Section: Football Datasetmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Holistic Human Pose Estimation with Regression Forests

Belagiannis

Amann

Navab

et al. 2014

Articulated Motion and Deformable Objects

View full text Add to dashboard Cite

show abstract

“…Pictorial structures is the most common part-based model for estimating the 2D body pose of single human [3,11,13]. The model has been extended to the 3D space, in order to cope with mutli-view camera setups as well [2,9,16]. Recently, pictorial structures have been successfully modelled for multiple human 3D pose estimation [6].…”

Section: Introductionmentioning

confidence: 99%

Multiple Human Pose Estimation with Temporally Consistent 3D Pictorial Structures

Belagiannis

Wang

Schiele

et al. 2015

Computer Vision - ECCV 2014 Workshops

View full text Add to dashboard Cite

Abstract. Multiple human 3D pose estimation from multiple camera views is a challenging task in unconstrained environments. Each individual has to be matched across each view and then the body pose has to be estimated. Additionally, the body pose of every individual changes in a consistent manner over time. To address these challenges, we propose a temporally consistent 3D Pictorial Structures model (3DPS) for multiple human pose estimation from multiple camera views. Our model builds on the 3D Pictorial Structures to introduce the notion of temporal consistency between the inferred body poses. We derive this property by relying on multi-view human tracking. Identifying each individual before inference significantly reduces the size of the state space and positively influences the performance as well. To evaluate our method, we use two challenging multiple human datasets in unconstrained environments. We compare our method with the state-of-the-art approaches and achieve better results.

show abstract

A Kinematic Chain Space for Monocular Motion Capture

Wandt

Ackermann

Rosenhahn

2019

Lecture Notes in Computer Science

View full text Add to dashboard Cite

This paper deals with motion capture of kinematic chains (e.g. human skeletons) from monocular image sequences taken by uncalibrated cameras. We present a method based on projecting an observation into a kinematic chain space (KCS). An optimization of the nuclear norm is proposed that implicitly enforces structural properties of the kinematic chain. Unlike other approaches our method does not require specific camera or object motion and is not relying on training data or previously determined constraints such as particular body lengths. The proposed algorithm is able to reconstruct scenes with limited camera motion and previously unseen motions. It is not only applicable to human skeletons but also to other kinematic chains for instance animals or industrial robots. We achieve state-of-the-art results on different benchmark data bases and real world scenes.

show abstract

Multi-view Body Part Recognition with Random Forests

Cited by 66 publications

References 19 publications

Holistic Human Pose Estimation with Regression Forests

Holistic Human Pose Estimation with Regression Forests

Multiple Human Pose Estimation with Temporally Consistent 3D Pictorial Structures

A Kinematic Chain Space for Monocular Motion Capture

Contact Info

Product

Resources

About