Organizers

• • A. Zanfir, E. Marinoiu, A.I. Popa, M. Zanfir, V. Olaru, C. Sminchisescu

Challenge

• • 3D Human Pose Estimation, within the ECCV 2018 PoseTrack Challenge

• • Create or sign-in to your account in order to participate!

• • Train a model that estimates the 3D joint positions of a single person in an image

• • Upload the results to see the performance on the withheld test-set

• • Check the Ranking for the CHALL_H80k testset to see how you compare to others

Rules

• • Maximum one submission per week.

• • Use of external 3D pose data for training is not allowed.

• • Only the 3D pose ground-truth will be provided.

• • Use of pretrained features on ImageNet (i.e. other tasks) is allowed.

• • Training on both train+val for a submission is allowed.

Challenge Dataset Structure

• • This is a subset of the large-scale dataset Human3.6M

• • 80,000 3D human poses and corresponding images

• • 10 professional actors (6 male, 4 female)

• • 15 scenarios (discussion, smoking, taking photo, talking on the phone...)

• • Click to View Train/Val Readme

• • Click to View Submission Readme

Important dates

• • Train+Val pre-release: July 9th, 2018

• • Full benchmark release and evaluation server opening: July 22nd, 2018

• • Results submission: August 18th, 2018

References

The datasets, large-scale learning techniques, and related experiments are described in:

• Catalin Ionescu, Dragos Papava, Vlad Olaru and Cristian Sminchisescu, Human3.6M: Large Scale Datasets and Predictive Methods for 3D Human Sensing in Natural Environments, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 36, No. 7, July 2014 [pdf][bibtex]

• Catalin Ionescu, Fuxin Li and Cristian Sminchisescu, Latent Structured Models for Human Pose Estimation, International Conference on Computer Vision, 2011 [pdf][bibtex]

The license agreement for data usage implies the citation of the two papers above. Please notice that citing the dataset URL instead of the publications would not be compliant with this license agreement.

Related Work

• Andrei Zanfir, Elisabeta Marinoiu, and Cristian Sminchisescu, Monocular 3D Pose and Shape Estimation of Multiple People in Natural Scenes The Importance of Multiple Scene Constraints, IEEE International Conference on Computer Vision and Pattern Recognition, 2018 [pdf]

• Mihai Zanfir, Alin-Ionut Popa, Andrei Zanfir and Cristian Sminchisescu, Human Appearance Transfer, IEEE International Conference on Computer Vision and Pattern Recognition, 2018 [pdf]

• Elisabeta Marinoiu, Mihai Zanfir, Vlad Olaru and Cristian Sminchisescu, 3D Human Sensing, Action and Emotion Recognition in Robot Assisted Therapy of Children with Autism, IEEE International Conference on Computer Vision and Pattern Recognition, 2018 [pdf]

• Alin-Ionut Popa, Mihai Zanfir and Cristian Sminchisescu, Deep Multitask Architecture for Integrated 2D and 3D Human Sensing, IEEE International Conference on Computer Vision and Pattern Recognition, 2017 [pdf]