2D Human Pose Estimation: New Benchmark and State of the Art Analysis

Andriluka, Mykhaylo; Pishchulin, Leonid; Gehler, Peter V.; Schiele, Bernt

doi:10.1109/cvpr.2014.471

Cited by 2,255 publications

(1,909 citation statements)

References 18 publications

Supporting

Mentioning

1,807

Contrasting

Unclassified

Order By: Relevance

“…In conjunction with the use of intermediate supervision, repeated bidirectional inference is critical to the network's final performance. The final network architecture achieves a significant improvement on the stateof-the-art for two standard pose estimation benchmarks (FLIC [1] and MPII Human Pose [21]). On MPII there is over a 2% average accuracy improvement across all joints, with as much as a 4-5% improvement on more difficult joints like the knees and ankles.…”

Section: Arxiv:160306937v2 [Cscv] 26 Jul 2016mentioning

confidence: 99%

“…We evaluate our network on two benchmark datasets, FLIC [1] and MPII Human Pose [21]. FLIC is composed of 5003 images (3987 training, 1016 testing) taken from films.…”

Section: Training Detailsmentioning

confidence: 99%

“…Recent pose estimation systems [15][16][17][18][19][20] have universally adopted ConvNets as their main building block, largely replacing hand-crafted features and graphical models; this strategy has yielded drastic improvements on standard benchmarks [1,21,22].…”

Section: Arxiv:160306937v2 [Cscv] 26 Jul 2016mentioning

confidence: 99%

See 2 more Smart Citations

Stacked Hourglass Networks for Human Pose Estimation

Newell

Yang

Deng

2016

Lecture Notes in Computer Science

4,087

4,081

View full text Add to dashboard Cite

Abstract. This work introduces a novel convolutional network architecture for the task of human pose estimation. Features are processed across all scales and consolidated to best capture the various spatial relationships associated with the body. We show how repeated bottom-up, top-down processing used in conjunction with intermediate supervision is critical to improving the performance of the network. We refer to the architecture as a "stacked hourglass" network based on the successive steps of pooling and upsampling that are done to produce a final set of predictions. State-of-the-art results are achieved on the FLIC and MPII benchmarks outcompeting all recent methods. Keywords: Human Pose Estimation

show abstract

Section: Arxiv:160306937v2 [Cscv] 26 Jul 2016mentioning

confidence: 99%

“…We evaluate our network on two benchmark datasets, FLIC [1] and MPII Human Pose [21]. FLIC is composed of 5003 images (3987 training, 1016 testing) taken from films.…”

Section: Training Detailsmentioning

confidence: 99%

See 1 more Smart Citation

Stacked Hourglass Networks for Human Pose Estimation

Newell

Yang

Deng

2016

Lecture Notes in Computer Science

4,087

4,081

View full text Add to dashboard Cite

show abstract

“…We perform extensive active learning experiments using the challenging MPII [1] and LSP datasets [19]. A first series of experiments using simulated annotators demonstrates that: (1) our proposed multiple peak entropy cue outperforms previous uncertainty-based cues; (2) our proposed dynamic combination of influence and uncertainty cues further improves active selection over individual cues and outperforms a static combination strategy.…”

Section: Introductionmentioning

confidence: 99%

“…These methods aim to learn discriminative patterns that enable to distinguish patches around body joints from the rest of the image. This requires good training data, but data collection is particularly time-intensive for human pose estimation, as annotators are typically asked to click on 14 joints per person [1]. The reference analysis paper [1] suggests a reasonable annotation rate of one pose per minute.…”

Section: Introductionmentioning

confidence: 99%

Active Learning for Human Pose Estimation

Liu

Ferrari

2017

2017 IEEE International Conference on Computer Vision (ICCV)

View full text Add to dashboard Cite

Annotating human poses in realistic scenes is very time consuming, yet necessary for training human pose estimators. We propose to address this problem in an active learning framework, which alternates between requesting the most useful annotations among a large set of unlabelled images, and re-training the pose estimator. To this end, (1) we propose an uncertainty estimator specific for body joint predictions, which takes into account the spatial distribution of the responses of the current pose estimator on the unlabelled images; (2) we propose a dynamic combination of influence and uncertainty cues, where their weights vary during the active learning process according to the reliability of the current pose estimator; (3) we introduce a computer assisted annotation interface, which reduces the time necessary for a human annotator to click on a joint by discretizing the image into regions generated by the current pose estimator. Experiments using the MPII and LSP datasets with both simulated and real annotators show that (1) the proposed active selection scheme outperforms several baselines; (2) our computer-assisted interface can further reduce annotation effort; and (3) our technique can further improve the performance of a pose estimator even when starting from an already strong one.

show abstract