Capturing 3D hand movements and reconstructing realistic hand models are crucial for real time hand object interaction in virtual/mixed reality environments. Recovering realistic hand poses is a challenging task due to the complexity of the hand shapes with 27 degrees of freedom, self occlusions and occlusions caused by physical objects. This project aims to tackle the problem of reconstructing a realistic hand mesh model with the texture of a human hand from a large dataset. The statistical models of the hands will be constructed. The ultimate mesh model must follow the correct geometrical and physical constraints of a real human hand. A large dataset of hands captured from various human subjects from mutiviews and the corresponding point cloud data is available in order to create the articulated non-rigid hand mesh model. The outcome must be compatible with standard graphics packages, and must fit any human hand shapes and poses.
ManoMotion is looking for a passionate and innovative intern or master thesis student in computer science, media technology, mathematics or relevant engineering subjects to work on the described project for a duration of three to six months. The expected outcome of the project will be implementations of the developed methods for realistic reconstruction of human hand model for real time interaction in XR applications.
- Master level education or similar level of experience in computer science, mathematics, or engineering subjects.
- Background knowledge in 3D graphics, 3D modeling, linear algebra, optimization, and statistics.
- Strong coding skills in Python or C++.
- Familiar with computer vision, computer graphics, animation and rendering concepts.
- Self-driven problem solver with a strong sense of teamwork.
ManoMotion is a Deep Tech company based in Stockholm, Sweden. We are a diverse team of Computer Vision and Machine Learning Scientists and creative software engineers, specialists in stretching the borders of what is technically possible to decipher from a camera. ManoMotion brings unparalleled intuition in human-machine interactions, refined through more than 12 years of research in gesture technology with a number of inventions and international awards. We have developed a core technology framework to achieve precise hand tracking and gesture recognition in 3D-space simply using a regular camera available on any smart device!