电子设计工程
電子設計工程
전자설계공정
Electronic Design Engineering
2015年
21期
15-17
,共3页
帧差法%滑动平均法%Viola-Jones算法%Boxfilter算法
幀差法%滑動平均法%Viola-Jones算法%Boxfilter算法
정차법%활동평균법%Viola-Jones산법%Boxfilter산법
frame difference method%moving average method%Viola-Jones algorithm%Boxfilter algorithm
为了进一步提高人脸检测的准确率，本文采用一种基于运动区域的检测方法，该方法主要通过采用帧差法和自适应的滑动平均法相结合的算法，建立提取运动区域的模型，然后通过基于Boxfilter算法的Viola-Jones进行快速人脸检测。自适应的滑动平均法是通过检测光照突变和像素点来实现阈值和更新率的自适应更新，该法提取的前景与帧差法提取的前景进行"与运算"。实验结果表明，该方法可以有效地提取运动区域，从而使人脸检测性能得到进一步的提高，具有一定的应用价值。
為瞭進一步提高人臉檢測的準確率，本文採用一種基于運動區域的檢測方法，該方法主要通過採用幀差法和自適應的滑動平均法相結閤的算法，建立提取運動區域的模型，然後通過基于Boxfilter算法的Viola-Jones進行快速人臉檢測。自適應的滑動平均法是通過檢測光照突變和像素點來實現閾值和更新率的自適應更新，該法提取的前景與幀差法提取的前景進行"與運算"。實驗結果錶明，該方法可以有效地提取運動區域，從而使人臉檢測性能得到進一步的提高，具有一定的應用價值。
위료진일보제고인검검측적준학솔，본문채용일충기우운동구역적검측방법，해방법주요통과채용정차법화자괄응적활동평균법상결합적산법，건립제취운동구역적모형，연후통과기우Boxfilter산법적Viola-Jones진행쾌속인검검측。자괄응적활동평균법시통과검측광조돌변화상소점래실현역치화경신솔적자괄응경신，해법제취적전경여정차법제취적전경진행"여운산"。실험결과표명，해방법가이유효지제취운동구역，종이사인검검측성능득도진일보적제고，구유일정적응용개치。
In order to improve the accuracy of face detection, this paper adopts a detection method in motion region. The method to extract moving regions model is established, through a algorithm of combining frame difference method and adaptive moving average method. Then faces are detected through Viola-Jones algorithm based on Boxfilter . The moving average method adoptes threshold and update rate by detecting the light adaptive mutation and the pixel. 'and operators’ is on the extracted foreground and that is extracted by frame difference method .The experimental results show that, this method can effectively extract moving regions. So that the face detection performance can be further improved, and it has certain application value.
