Electronics and Telecommunications Research Institute

Jong Sung Kim

Seong Min Baek

Myung Gyu Kim

Seong Il Yang

Sang Woo Seo

Sung Jin Hong

Il Kwon Jeong

Min Sung Yoon

Woo Suk Kim

Hyuk Jeong

Youn Hee Gil

Jae-Hean Kim

Wookho Son

Byoung Tae Choi

Si Hwan Jang

Hye-Mi Kim

Jung Jae Yu

Jaehwan Kim

Hee Sook Shin

Jin Ho Kim

Kwang Ho Yang

Cho Rong Yu

Il-Kwon Jeong

Seung Joon Kwon

Dae Wook Kim

Yeong Jae Choi

Minsung Yoon

Kwang-Hyun Shim

Kyung Jun Song

Kwang-Man Oh

Byung Eui Min

A method and an apparatus for registering motion of a user tracked by heterogeneous sensors are disclosed. The method includes initializing the first sensor so that an origin of a first sensor coordinate system is adjacent to an origin of a device coordinate system, wherein the device coordinate system is a three-dimensional coordinate system with a current position of the device as the origin; and determining a calibration parameter for transforming coordinates of the plurality of feature points according to the first sensor coordinate system to coordinates according to a second sensor coordinate system, based on the position and orientation of the device in the second sensor coordinate system.

Method and apparatus for matching user motion tracked by heterogeneous sensors

The present disclosure provides a method and device for analyzing user satisfaction of screen sports contents. According to one embodiment, the present disclosure provides a method of analyzing user satisfaction of a screen sports content, including generating first satisfaction detection information based on real-time chat data of a user using a text feature recognition model, generating second satisfaction detection information based on real-time chat data of the user using a language feature recognition model, and predicting real-time satisfaction of the user based on the first satisfaction detection information and the second satisfaction detection information using a satisfaction prediction model.

Method and apparatus for analyzing satisfaction of screen sports contents user

A method of synthesizing 3D joint data may include converting joint data collected from a depth camera of each of a plurality of RGB-D cameras from a depth camera coordinate system of each of the RGB-D cameras to a color camera coordinate system of each of the RGB-D cameras, and calculating a confidence level of the joint data converted to the color camera coordinate system using a 3D joint recognition algorithm based on the joint data converted to the color camera coordinate system. The method may further include applying a rotation matrix and a translation vector to the joint data and converting the joint data to a predetermined reference coordinate system, and obtaining a weighted-average of the joint data converted to the reference coordinate system using a weight calculated based on the confidence level to synthesize the joint data.

Method of synthesizing 3D joint data based on multi-view RGB-D camera

Provided is an impact motion recognition system for screen-based multi-sport coaching. The impact motion recognition system for screen-based multi-sport coaching includes: a high-speed stereo camera device installed on an upper end of a target object including a hitting body and an moving body and configured to acquire a stereo image sequence in which the hitting body and the moving body are photographed; an object detection unit configured to detect an object in the acquired stereo image sequence; a moving body recognition unit configured to identify whether an impact motion has occurred in the detected object to recognize the moving body; a hitting body detection unit configured to, based on the moving body being recognized through the moving body recognition unit, detect the object excluding the moving body as the hitting body; a target object tracking unit configured to continuously detect a center point of the moving body and a feature point of the hitting body in a subsequent stereo image sequence using characteristics of the moving body and the hitting body that are detected, and continuously track the center point of the moving body and the feature point of the hitting body until the center point of the moving body and the feature point of the hitting body are not detected; and a motion analysis unit configured to analyze three-dimensional (3D) motion information of the target object using the center point of the moving body and the feature point of the hitting body, which are detected through the target object tracking unit, and high-speed stereo camera correction data.

Impact motion recognition system for screen-based multi-sport coaching

Provided are an apparatus and method for experiencing an augmented reality (AR)-based screen sports match which enable even a child, an elderly person, and a person with a disability to easily and safely experience a ball sports match, such as tennis, badminton, or squash, as a screen sports match without using a wearable marker or sensor, a wearable display, an actual ball, and an actual tool.

Apparatus and method for experiencing augmented reality-based screen sports match

Provided is a system and method for automatically recognizing user motion. The system for automatically recognizing user motion includes an input unit configured to receive three-dimensional (3D) measurement data, a memory which stores a program for performing automatic recognition on 3D user motion using 3D low-quality depth data and a deep learning model, and a processor configured to execute the program, wherein the processor converts the 3D low-quality depth data into 3D high-quality image data.

System and method for automatic recognition of user motion

A system for recognizing collision positions of a plurality of moving objects in a screen includes an infrared camera configured to obtain an image frame when a plurality of moving objects move toward a screen, a memory configured to store a program for calculating a collision position in the screen on the basis of the image frame, and a processor configured to execute the program stored in the memory. By executing the program, the processor detects all of the plurality of moving objects from the image frame to generate a graph where moving objects included in a successive image frame are connected to one another, and calculates a collision position of the moving object included in the image frame and a collision position of the moving object in the screen on the basis of the graph.

System and method of recognizing collision position of screen

Provided are a system and method for simultaneously reconstructing an initial three-dimensional (3D) trajectory and velocity of an object by using single camera images. The system for simultaneously reconstructing an initial 3D trajectory and velocity of an object by using single camera images includes a receiver configured to receive 3D world coordinate information of a starting point of an object and object-captured image information of a single camera, a two-dimensional (2D) coordinate acquisition unit configured to acquire 2D image coordinates of the object from the object-captured image information, and a reconstructor configured to simultaneously reconstruct an initial 3D velocity and trajectory of the object by applying a 3D motion equation to the 2D image coordinates and depth information from the single camera to the object.

System and method for simultaneously reconstructing initial 3D trajectory and velocity of object by using single camera images

Provided is an apparatus and method for automatically recognizing an object by using a low-speed camera in a dual photographing mode, in which the apparatus includes an image acquirer and an image information analyzer. In the present disclosure, without using a separate motion sensor a movement state of a specific object during sports activities may be automatically recognized by using a successive photographing function of a low-speed camera in a dual photographing mode. Further, once an object starts moving, high-speed images in a multiple exposure mode may be automatically captured by using a high-speed photographing function of the low-speed camera in a dual photographing mode without using an expensive high-speed camera, and based on the captured images, movement information, including a movement speed and direction of the sports ball, the force and axis of rotation of the ball, and the like, is automatically analyzed to generate object movement information.

Apparatus and method for automatically recognizing object by using low-speed camera in dual photographing mode

Disclosed are a method and apparatus for measuring rotation characteristics such as rotation rate, rotation axis, and rotation angle of a rotating body. A method of measuring rotation characteristics includes extracting a rotating-body region from an image of the rotating body; extracting a surface pattern of the rotating body on the basis of brightness values of the extracted rotating-body region and acquiring rotation characteristics of the rotating body on the basis of change in the extracted surface pattern. Accordingly, it is possible to accurately measure rotation characteristics regardless of variation in brightness values of a surface region of the rotating body depending on illumination of a lamp, sensitivity of a camera, and exposure time of a camera.

Method and apparatus for measuring rotation characteristics of rotating body

Disclosed are an apparatus and a method for measuring a speed of a rotation body and recognizing a spin using a line scan. The present invention provides an apparatus and a method for calculating a motion of an object capable of acquiring line scan images using some lines of an area scan camera and calculating a three-dimensional speed and a three-dimensional spin of a rotation body by using a composite image in which the line scan images are coupled. The present invention can provide a realistic game or training at low cost by providing realistic physical simulation of the rotation body while allowing a competitive price when producing products by using the existing inexpensive camera.

Apparatus and method for calculating motion of object

Disclosed is a method for automatically extracting centroids and features of a spherical object required to measure a flight speed, a flight direction, a rotation speed, and a rotation axis of the spherical object in a system for measuring flight information of the spherical object with a high-speed stereo camera. In order to automatically extract the centroids and the features of the spherical object, the present invention may automatically extract the centroid and the feature by detecting only the pixel of the foreground image including the spherical object generated by excluding the motionless background image from each camera image, sorting the interconnected pixels among the pixels of the detected foreground image into the independent pixel cluster, and then, using, in the extraction of the centroid and the feature, only one pixel cluster having a size similar to the actual spherical object.

System and method for measuring flight information of a spherical object with high-speed stereo camera

An apparatus for calibrating images between cameras, includes: a detector for detecting dynamic markers from images taken by a motion capture camera and a video camera; a 3D position recovery unit for recovering 3D position data of the dynamic markers obtained by the motion capture camera; a 2D position tracking unit for tracking 2D position data of the dynamic markers obtained by the video camera; a matching unit for matching the 3D position data and the 2D position data of the dynamic markers; and a calibrating unit for converting the 3D position of the dynamic markers on a spatial coordinate system into image coordinates on an image coordinate system of the video camera to calculate intrinsic parameters and extrinsic parameters of the video camera to minimize a difference between coordinates of the 2D position data and the image coordinates.

Apparatus and method for calibrating images between cameras

A method for restoration of building structure using infinity homographies calculated based on parallelograms includes: calculating, using two or more parallelograms, an infinity homography between those cameras which refer to an arbitrary camera; restoring cameras and the building structure on an affine space using the computed infinity homography and homologous points between images; and transforming the restored result onto the metric space using constraints on orthogonality of vectors joining the restored three-dimensional points, the ratio of lengths of the vectors and intrinsic camera parameters. As a result, intrinsic camera parameters, camera positions on the metric space and the structure of the building are restored. All the restoration is possible even when intrinsic camera parameters corresponding to all the images are not constant.

Method for restoration of building structure using infinity homographies calculated based on parallelograms

An apparatus and method for controlling vehicle by teeth clenching are provided. The apparatus includes: an electromyogram signal obtaining unit including electromyogram sensors disposed at both sides for generating an electromyogram signal according to a predetermined muscle moved when a disabled person clenches teeth, and a ground electrode connected to a body of the disabled persons for providing a reference voltage; and a vehicle driving unit including a control command generating unit for generating a vehicle driving command according to the electromyogram signal by classifying the electromyogram signal based on a side of teeth clenched, a duration time for clenching teeth and a sequence of teeth clenching motions made by the disabled person, a control command interface for generating a predetermined level of voltage according to the created vehicle driving command, and a vehicle driving unit for driving the vehicle according to the generated voltage.

Growing community of inventors

Daejeon, South Korea

Jong Sung Kim