公开(公告)号：KR100399762B1

公开(公告)日：2003-09-29

申请号：KR1020010069711

申请日：2001-11-09

Applicant: 한국과학기술원

Inventor： 임종태 ,  최호림 ,  변희정 ,  송원규 ,  손준원

IPC: A61B5/04

Abstract: PURPOSE: A real-time pattern learning and classifying method is provided to be capable of finding start and end points of a signal by dividing an entire electromyogram signal into small periods and giving an energy reference every divided signal according to a signal magnitude. CONSTITUTION: An entire electromyogram is divided into predetermined number of small periods, an energy level is calculated by the first equation every period, and a specific energy value is calculated by the second equation using the calculated energy level. Predetermined number of threshold values is obtained by the calculated energy value. Increase in an energy level due to a noise is seized by judging whether the obtained energy level satisfies the first and second conditions. If not, an energy reference point is reset to '0', whether the obtained energy level satisfies the third condition is judged while sequentially increasing the divided period by '1'. If the obtained energy level satisfies the third condition, whether the obtained energy level reaches the last energy reference point is judged while increasing an energy reference point by '1'. Judging whether the obtained energy level reaches the last energy reference point, is repeated.

Abstract translation: 目的：提供一种实时模式学习和分类方法，能够通过将整个肌电信号分成小周期并根据信号幅度给每个分割信号提供能量基准来找到信号的起点和终点。 结构：整个肌电图被分成预定数量的小周期，每个周期通过第一个方程计算能量水平，并且使用计算的能量水平通过第二个方程计算特定能量值。 通过计算的能量值获得预定数量的阈值。 通过判断获得的能量水平是否满足第一和第二条件来检测由于噪声导致的能量水平的增加。 如果不是，则将能量参考点重置为'0'，同时依次将分频周期增加'1'来判断所获得的能量水平是否满足第三条件。 如果所获得的能量水平满足第三条件，则在将能量参考点增加'1'的同时判断所获得的能量水平是否达到最后的能量参考点。 判断获得的能量水平是否达到最后的能量参考点，是重复的。

公开(公告)号：KR1020030038935A

公开(公告)日：2003-05-17

申请号：KR1020010069711

申请日：2001-11-09

Applicant: 한국과학기술원

Inventor： 임종태 ,  최호림 ,  변희정 ,  송원규 ,  손준원

IPC: A61B5/04

Abstract: PURPOSE: A real-time pattern learning and classifying method is provided to be capable of finding start and end points of a signal by dividing an entire electromyogram signal into small periods and giving an energy reference every divided signal according to a signal magnitude. CONSTITUTION: An entire electromyogram is divided into predetermined number of small periods, an energy level is calculated by the first equation every period, and a specific energy value is calculated by the second equation using the calculated energy level. Predetermined number of threshold values is obtained by the calculated energy value. Increase in an energy level due to a noise is seized by judging whether the obtained energy level satisfies the first and second conditions. If not, an energy reference point is reset to '0', whether the obtained energy level satisfies the third condition is judged while sequentially increasing the divided period by '1'. If the obtained energy level satisfies the third condition, whether the obtained energy level reaches the last energy reference point is judged while increasing an energy reference point by '1'. Judging whether the obtained energy level reaches the last energy reference point, is repeated.

Abstract translation: 目的：提供一种实时模式学习和分类方法，以便通过将整个肌电图信号分成小周期并根据信号幅度给出每个分频信号的能量基准来找到信号的起点和终点。 构成：将整个肌电图划分成预定数量的小周期，每周期通过第一方程计算能量水平，并且使用计算的能量级通过第二方程计算比能量值。 通过计算的能量值获得预定数量的阈值。 通过判断所获得的能量水平是否满足第一和第二条件，来检测由于噪声引起的能级的增加。 如果不是，则能量参考点被重置为“0”，判断所获得的能级是否满足第三条件，同时依次将分割周期增加1。 如果获得的能级满足第三条件，则在将能量基准点增加1的同时判断所获得的能级达到最后的能量基准点。 判断所获得的能量达到最后的能量参考点是否重复。