The 3D motion capture system is a technology that records the movement of people or objects, which is subsequently changed into actionable data used to create a 3D view of the whole performance. The system is applicable for several applications, such as sports, robotics, ergonomics, medical, and entertainment.
In a significant step towards advancing the same technology, researchers have now developed a new real-time, 3D Motion Tracking System that combines advanced neural network methods with transparent light detectors. This can be considered massive progress for 3D Motion Capture System Market as the system is expected to replace LiDAR (Light Detection and Ranging) and other cameras in autonomous technologies. Moreover, the prototype made might bring benefits for the autonomous driving and robotics sector. The system can respond to moving objects by determining the what, where, and how far the object is in real-time.
The photodetector array works by measuring the focal stack images of a point object. This is achieved by emphasizing a free laser beam into a small spot ahead of the lens. The team constructed a prototype of transparent photodetector arrays which had graphene on glass. They then used the two pieces of the detector arrays to prove the potential of the 3D object tracking tasks.
The novel imaging system is one of a kind and utilizes the advantages of highly sensitive, transparent, nanoscale graphene photodetectors. It brings together computational power efficiency, compact hardware, fast-tracking speed, and a lower cost compared to other approaches present in the market. Furthermore, the in-depth grouping of machine learning algorithms and graphene nanodevices might bring forth intriguing opportunities for both science and technological sectors.
In the study, graphene photodetectors were changed to absorb at most 10% of the light to which it is exposed. This gives them the ability to be transparent. As graphene is known to be sensitive to light, this amount of light is enough to produce images that can recreate with the help of computational imaging. Furthermore, the photodetectors are placed one after the other, leading to a compact system, in which each layer specializes on a different focal plane, which facilitates 3D imaging.
The team taught the neural network to search for specific objects around the entire scene. Then, it is made to focus on a particular object that is of interest. For instance, an object suddenly comes in front, in a lane on a highway, or a pedestrian in traffic. The technology is advantageous for stable systems, such as projection of human body structures or automated manufacturing in 3D for medical study and research.
This developed algorithm is mainly different from traditional signal processing algorithms like MRI (Magnetic Resonance Imaging) and X-ray. The technology has several advantages and can be used with other materials as well. However, its focal aspect is that it does not require artificial illumination and is extremely environment friendly. The team stated that their research needs further work as building manufacturing infrastructure for industrial production of this approach would be an arduous task for now.