微型机与应用
微型機與應用
미형궤여응용
MICROCOMPUTER & ITS APPLICATIONS
2012年
14期
40-43
,共4页
人脸检测%DM6437%Adaboost%OpenCV
人臉檢測%DM6437%Adaboost%OpenCV
인검검측%DM6437%Adaboost%OpenCV
face detection%DM6437%Adaboost%OpenCV
研究并实现了基于DM6437的Adaboost人脸检测算法。在对相关的人脸检测算法研究的基础上,选择了适应能力强、错误率小的Adaboost算法,通过对输入样本进行Harr特征提取,从中选出最优的Haar特征,然后将训练得到的Haar特征转换成弱分类器,再将弱分类器优化组合成强分类器,最后形成级联强分类器用于人脸检测。通过OpenCV在计算机上仿真实现该算法’,完成了Adaboost人脸检测算法的DSP程序设计,在DM6437硬件平台上实现了人脸实时检测功能。结果表明,运用该算法能够有效地进行人脸检测,可用于工程实践。
研究併實現瞭基于DM6437的Adaboost人臉檢測算法。在對相關的人臉檢測算法研究的基礎上,選擇瞭適應能力彊、錯誤率小的Adaboost算法,通過對輸入樣本進行Harr特徵提取,從中選齣最優的Haar特徵,然後將訓練得到的Haar特徵轉換成弱分類器,再將弱分類器優化組閤成彊分類器,最後形成級聯彊分類器用于人臉檢測。通過OpenCV在計算機上倣真實現該算法’,完成瞭Adaboost人臉檢測算法的DSP程序設計,在DM6437硬件平檯上實現瞭人臉實時檢測功能。結果錶明,運用該算法能夠有效地進行人臉檢測,可用于工程實踐。
연구병실현료기우DM6437적Adaboost인검검측산법。재대상관적인검검측산법연구적기출상,선택료괄응능력강、착오솔소적Adaboost산법,통과대수입양본진행Harr특정제취,종중선출최우적Haar특정,연후장훈련득도적Haar특정전환성약분류기,재장약분류기우화조합성강분류기,최후형성급련강분류기용우인검검측。통과OpenCV재계산궤상방진실현해산법’,완성료Adaboost인검검측산법적DSP정서설계,재DM6437경건평태상실현료인검실시검측공능。결과표명,운용해산법능구유효지진행인검검측,가용우공정실천。
The paper studies realizes the Adaboost face detection algorithm based on DM6437. After studying the relevant algo- rithms, the paper chooses Adaboost face detection algorithm which is adaptable and has small error rate. Firstly ,it extracts the Hart features of input samples and chooses optimal Harr features through training, and changes it into weak classifier. Then, it makes weak classifier optimized into strong classifier. Finally it forms cascade classifier which is used in face detection. It simulates Ad- aboost face detection algorithm through OpenCV on the computer. The paper completes the program design of DSP and realizes the function of real-time face detection in DM6437 hardware platform. Results show that this algorithm can be used to realize face de- tection efficiendy, and can be used in engineering practice.