Abstract:
An apparatus and method for generating synthetic training data for motion recognition. The method includes generating a three-dimensional (3D) human body model in real time according to motion of a human body, converting volume data of the 3D human body model into a 3D linear structure and extracting skeletal joint information, generating a data set of a human body image and skeletal joint information by rendering the 3D human body model and the skeletal joint information, and providing the data set of the image and the skeletal joint information as synthetic training data for motion recognition.
Abstract:
An apparatus for 3D reconstruction based on multiple GPUs and a method thereof are disclosed. The 3D reconstruction apparatus according to the present invention includes a 3D reconstruction apparatus, comprising: a camera configured to generate depth data for 3D space; a first GPU configured to update first TSDF volume data with first depth data generated for a first area and predict a surface point of an object which is present in the space from the first updated TSDF volume data; a second GPU configured to update second TSDF volume data with second depth data generated for a second area and predict a surface point of an object which is present in the space from the second updated TSDF volume data; and a master GPU configured to combine a surface point predicted from the first TSDF volume data and a surface point estimated from the second TSDF volume data.
Abstract:
Disclosed are a charge sensitive amplifier capable of minimizing a variation in a signal voltage of an output signal by applying a bias direct current to a gate of a feedback transistor, and a radiation sensor including the same. According to the charge sensitive amplifier and the radiation sensor including the same, it is possible to minimize a variation in a signal voltage of a charge sensitive amplifier output signal by applying a current, which is formed by mirroring a current bias circuit designed to be insensitive to PVT variations, to a gate of a feedback transistor. Furthermore, it is possible to reduce a variation in charging time and enable high-speed sensing by charging the signal voltage to the level of a common voltage VCOM by using a constant current supplied through a bandgap reference (BGR) circuit.
Abstract:
Provided are an apparatus and method for tracking a camera that reconstructs a real environment in three dimensions by using reconstruction segments and a volumetric surface. The camera tracking apparatus using reconstruction segments and a volumetric surface includes a reconstruction segment division unit configured to divide three-dimensional space reconstruction segments extracted from an image acquired by a camera, a transformation matrix generation unit configured to generate a transformation matrix for at least one reconstruction segment among the reconstruction segments obtained by the reconstruction segment division unit, and a reconstruction segment connection unit configured to rotate or move the at least one reconstruction segment according to the transformation matrix generated by the reconstruction segment division unit and connect the rotated and moved reconstruction segment with another reconstruction segment.