Holistic Human Pose Estimation with Regression Forests

Belagiannis, Vasileios; Amann, Christian; Navab, Nassir; Ilić, Slobodan

doi:10.1007/978-3-319-08849-5_3

Cited by 24 publications

(21 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Increasing the variability for a few iterations helps the network to quickly reach a more stable state. Note that we have empirically observed that the number of iterations needed In all datasets (PARSE [48], LSP [19], Football [20] and Volleyball [3]), T ukey s biweight loss function shows, on average, faster convergence and better generalization than L2. Both loss functions are visualised for the same number of epochs.…”

Section: Training Detailsmentioning

confidence: 94%

“…We evaluate Tukey's biweight loss function for the problem of 2D human pose estimation from still images. For that purpose, we have selected four publicly available datasets, namely PARSE [48], LSP [19], Football [20] and Volleyball [3]. All four datasets include sufficient amount of data for training the ConvNets, except for PARSE which has only 100 training images.…”

Section: Methodsmentioning

confidence: 99%

“…8). [3] is the loose PCP score. Table 1: Comparison with related work: We compare our results (Tukey's biweight loss) using one ConvNet (second row) and the cascade of ConvNets (third row).…”

Section: Comparison With Other Methodsmentioning

confidence: 99%

See 2 more Smart Citations

Robust Optimization for Deep Regression

Belagiannis

Rupprecht

Carneiro

et al. 2015

2015 IEEE International Conference on Computer Vision (ICCV)

Self Cite

171

137

View full text Add to dashboard Cite

Convolutional Neural Networks (ConvNets) have successfully contributed to improve the accuracy of regressionbased methods for computer vision tasks such as human pose estimation, landmark localization, and object detection. The network optimization has been usually performed with L2 loss and without considering the impact of outliers on the training process, where an outlier in this context is defined by a sample estimation that lies at an abnormal distance from the other training sample estimations in the objective space. In this work, we propose a regression model with ConvNets that achieves robustness to such outliers by minimizing Tukey's biweight function, an M-estimator robust to outliers, as the loss function for the ConvNet. In addition to the robust loss, we introduce a coarse-to-fine model, which processes input images of progressively higher resolutions for improving the accuracy of the regressed values. In our experiments, we demonstrate faster convergence and better generalization of our robust loss function for the tasks of human pose estimation and age estimation from face images. We also show that the combination of the robust loss function with the coarse-to-fine model produces comparable or better results than current state-of-the-art approaches in four publicly available human pose estimation datasets.

show abstract

Section: Training Detailsmentioning

confidence: 94%

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Robust Optimization for Deep Regression

Belagiannis

Rupprecht

Carneiro

et al. 2015

2015 IEEE International Conference on Computer Vision (ICCV)

Self Cite

171

137

View full text Add to dashboard Cite

show abstract

“…Randomized trees [165] and random forests [166,167] are fast and robust classification techniques that can handle this type of problem [266]. Random Forest is an ensemble classifier that consists of several randomized decision trees [142,267] and has a nonterminal node containing a decision function to predict the correspondences by regressing from images to terminal nodes, like mesh vertices [9] (Figure 14 shows an example). Enhanced random forests were used by [268], which employed two-layered random forests as joint regressors, with the first layer acting as a discriminative body part classifier and the second one predicting joint locations according to the results of the first layer.…”

Section: Methodologiesmentioning

confidence: 99%

Human Pose Estimation from Monocular Images: A Comprehensive Survey

Gong

Zhang

Gonzàlez

et al. 2016

Sensors

130

View full text Add to dashboard Cite

Human pose estimation refers to the estimation of the location of body parts and how they are connected in an image. Human pose estimation from monocular images has wide applications (e.g., image indexing). Several surveys on human pose estimation can be found in the literature, but they focus on a certain category; for example, model-based approaches or human motion analysis, etc. As far as we know, an overall review of this problem domain has yet to be provided. Furthermore, recent advancements based on deep learning have brought novel algorithms for this problem. In this paper, a comprehensive survey of human pose estimation from monocular images is carried out including milestone works and recent advancements. Based on one standard pipeline for the solution of computer vision problems, this survey splits the problem into several modules: feature extraction and description, human body models, and modeling methods. Problem modeling methods are approached based on two means of categorization in this survey. One way to categorize includes top-down and bottom-up methods, and another way includes generative and discriminative methods. Considering the fact that one direct application of human pose estimation is to provide initialization for automatic video surveillance, there are additional sections for motion-related methods in all modules: motion features, motion models, and motion-based methods. Finally, the paper also collects 26 publicly available data sets for validation and provides error measurement methods that are frequently used.

show abstract

“…By contrast, we choose to apply our model in the operating room scenario which comprises significantly mutual occlusions between the target individuals and thus is much more challenging than the normal ones. In contrary to the concept of part-based models, the holistic models predict directly the body pose by learning a mapping between features and poses [1,6,23,25,29,44,47,59]. One very popular method to accomplish this task are random forests for human pose estimation from depth data [22,42].…”

Section: Related Workmentioning

confidence: 99%

Parsing human skeletons in an operating room

Belagiannis

Wang

Shitrit

et al. 2016

Machine Vision and Applications

Self Cite

View full text Add to dashboard Cite

Multiple human pose estimation is an important yet challenging problem. In an Operating Room (OR) environment, the 3D body poses of surgeons and medical staff can provide important clues for surgical workflow analysis. For that purpose, we propose an algorithm for localising and recovering body poses of multiple human in an OR environment under a multi-camera setup. Our model builds on 3D Pictorial Structures (3DPS) and 2D body part localization across all camera views, using Convolutional Neural Networks (ConvNets). To evaluate our algorithm, we introduce a dataset captured in a real OR environment. Our dataset is unique, challenging and publicly available with annotated ground truths. Our proposed algorithm yields to promising pose estimation results on this dataset.

show abstract

Holistic Human Pose Estimation with Regression Forests

Cited by 24 publications

References 35 publications

Robust Optimization for Deep Regression

Robust Optimization for Deep Regression

Human Pose Estimation from Monocular Images: A Comprehensive Survey

Parsing human skeletons in an operating room

Contact Info

Product

Resources

About