Modeling of human motion is a long-standing problem in Computer Science and a central aspect in many related research fields such as Computer Graphics, Computer Vision, Robotics, or Biomechanics. Human motion is inherently complex because it is non-linear, high-dimensional and subject to great variability. Furthermore, we as humans are exceptionally skilled at detecting flaws and irregularities in human motion, which sets a high bar for any computational method producing such motion. Hence, modeling human motion requires considering both highly correlated nature of the data and stochasticity in the data.
In this project, we would like to design a generative neural network model leveraging temporal information and considering kinematic chain of the human body. Firstly, the model should be able to synthesize novel motion samples. Secondly, we can extend the model to allow conditional generation. In other words, one can generate new motion samples of given type such as _walking_, _bending_ or _sitting_. Such a model with high level control can also be employed in developing advanced animation design tools.
The ideal candidate for this position should have some background in deep learning. An existing code base in Tensorflow to train and evaluate models will be provided.
Modeling of human motion is a long-standing problem in Computer Science and a central aspect in many related research fields such as Computer Graphics, Computer Vision, Robotics, or Biomechanics. Human motion is inherently complex because it is non-linear, high-dimensional and subject to great variability. Furthermore, we as humans are exceptionally skilled at detecting flaws and irregularities in human motion, which sets a high bar for any computational method producing such motion. Hence, modeling human motion requires considering both highly correlated nature of the data and stochasticity in the data.
In this project, we would like to design a generative neural network model leveraging temporal information and considering kinematic chain of the human body. Firstly, the model should be able to synthesize novel motion samples. Secondly, we can extend the model to allow conditional generation. In other words, one can generate new motion samples of given type such as _walking_, _bending_ or _sitting_. Such a model with high level control can also be employed in developing advanced animation design tools.
The ideal candidate for this position should have some background in deep learning. An existing code base in Tensorflow to train and evaluate models will be provided.
Each year the IDEA League offers the students of its partner universities over 180 monthly grants for a short-term research exchange. In general, these grants are awarded based on academic merit. For more information visit http://idealeague.org/student-grant/