公开(公告)号：KR20090016800A

公开(公告)日：2009-02-18

申请号：KR20070081049

申请日：2007-08-13

Applicant: KOREA ELECTRONICS TELECOMM

Inventor： KIM HYE JIN ,  LEE JAE YEON ,  BAE KYUNG SOOK ,  KWAK KEUN CHANG ,  YOON HO SUB

IPC: G10L15/30 ,  G10L15/06 ,  G10L15/22

Abstract: A voice processing method and a system thereof are provided to control plural clients through one central server on basis of wired/wireless networks and integrally perform voice-related-application programs for robot. A sound is inputted through plural microphones and sound board(360). The inputted sound data is stored in a frame buffer, and IDs are allocated to each frame. The sound data of the same contents by ID is temporarily stored in a stack of the same location. The voice-related-application programs of robot are driven through the synchronized sound data. The microphone is a directional microphone(306) or non-directional multi array microphone(304).

Abstract translation: 提供语音处理方法及其系统，以有线/无线网络为基础通过一个中央服务器来控制多个客户端，并且整体地执行机器人的语音相关应用程序。 通过多个麦克风和声音板（360）输入声音。 输入的声音数据被存储在帧缓冲器中，并且将ID分配给每个帧。 通过ID的相同内容的声音数据被临时存储在相同位置的堆叠中。 机器人的语音相关应用程序通过同步声音数据驱动。 麦克风是定向麦克风（306）或非定向多阵列麦克风（304）。

公开(公告)号：KR100776801B1

公开(公告)日：2007-11-19

申请号：KR20060067215

申请日：2006-07-19

Applicant: KOREA ELECTRONICS TELECOMM

Inventor： YOON HO SUB ,  CHI SU YOUNG ,  KIM DO HYUNG ,  KIM HYE JIN ,  LEE JAE YEON

IPC: G06K9/80 ,  G06K9/36

CPC classification number: G06K9/00355 ,  G06K9/00228

Abstract: An apparatus for recognizing a gesture in a picture processing system and a method thereof are provided to be useful for the old or the physically handicapped by generating a control signal via recognition of the gesture and controlling video devices like a TV, a home robot or a game device via a simple motion without using a remote controller. A method for recognizing a gesture in a picture processing system comprises the following several steps. The picture processing system detects a face section from a picture captured and inputted by a camera(310,320). The picture processing system divides the inputted picture into plural gesture search sections on the basis of a set position of the detected face section. The picture processing system detects a hand section from the plural gesture search sections by using skin colour information obtained from the detected face section(330). The picture processing system checks whether a gesture in each gesture search section is included by using an image of the hand section via the skin colour information and an image of movement difference(360). The picture processing system transmits a predetermined control command in correspondence with a gesture search section where a gesture is sensed(390).

Abstract translation: 一种用于识别图像处理系统中的手势的装置及其方法被提供为通过经由识别手势并且控制诸如TV，家庭机器人或家庭机器人的视频设备来生成控制信号而对旧的或身体残疾者有用 游戏设备通过简单的运动而不使用遥控器。 一种在图像处理系统中识别手势的方法包括以下几个步骤。 图像处理系统从由照相机捕获并输入的图像（310,320）中检测人脸部分。 图像处理系统基于检测到的脸部的设定位置将输入的图像分割为多个手势搜索部。 图像处理系统通过使用从检测到的脸部部分（330）获得的肤色信息来检测来自多个手势搜索部分的手部。 图像处理系统经由皮肤颜色信息和运动差异图像（360）检查手势搜索部分中的手势是否包括手部分的图像。 图像处理系统与手势检测部分对应地发送预定的控制命令（390）。