Deep Learning Robotics (DLR), a leading innovator in the field of robotics and artificial intelligence, announced at the AI Week in Tel-Aviv the grant of a new patent for a cutting-edge technology enabling the training of objects' visual and 3D features to computer vision models with the assistance of robotic devices.
It is clear today that high-quality training data is crucial for computer vision deep learning models and AI. Back in 2016 we pioneered the usage of robotic devices to create such data by generating and linking visual and spatial features of objects automatically and in real-time, leading to more accurate models and faster deployment to robotic vision systems.
According to Carlos Benaim, the CEO of DLR, this novel technology has the potential to revolutionize the existing techniques of generating 3D data for training computer vision models. Furthermore, it will facilitate the creation of robot applications capable of accurately identifying and locating objects in a given scene, even when using basic 2D cameras.
DLR is committed to pushing the boundaries of what is possible with robotics and artificial intelligence, and this new patent is a testament to that commitment. The company is excited to continue to explore the full potential of this technology and bring new solutions to market in the years to come.