Pictorial Human Spaces

How Well do Humans Perceive a 3D Articulated Pose?

Elisabeta Marinoiu     Dragos Papava     Cristian Sminchisescu

Human motion analysis in images and video is a central computer vision problem. Yet, there are no studies that reveal how humans perceive other people in images and how accurate they are. We aim to unveil some of the processing as well as the levels of accuracy involved in the 3D perception of people from images by assessing the human performance.

The dataset, the analysis on human 3D pose re-enactment, and related experiments are described in detail in:

Elisabeta Marinoiu, Dragos Papava and Cristian Sminchisescu Pictorial Human Spaces: How Well do Humans Perceive a 3D Articulated Pose?, International Conference on Computer Vision (ICCV), 2013 [pdf ][bibtex ]

Elisabeta Marinoiu, Dragos Papava and Cristian Sminchisescu, Pictorial Human Spaces: A Computational Study on the Human perception of 3D Articulated Poses International Journal of Computer Vision (IJCV), 2016 [pdf ][ bibtex ]

3D Human Motion Capture System

SMI Head Mounted Eyetracking System

  • 200Hz sample rate

  • 50fps scene video data

  • 10 high speed motion camers

  • 200Hz eyetracking sample rate

  • 25fps scene video data

Data download



Read Me

Readme file explaining the structure of the dataset.

Pose Images

Images shown to subjects.

Pose Ground Truth

2D and 3D joint positions for each of 120 poses shown to subjects.

Mocap Data

Motion capture data for both regular and skilled subjects during pose re-enactment.

Eye Movements

Eye movements from regular and skilled subjects mapped to the original images.

Synchronized Frames

Synchronization between the Eye Tracker and Motion Capture systems.

For further information please contact the authors at the following email address: