| Face detection is to determine whether there existsin a picture or continuous video ,if have, marked the size and location of the face.It is one of the most popular comput-er pattern recognition research topics,is applied in video conference, remote monitoring, remote treatment. Many scholars,and experts have done a lot a lot of work and achieved outstanding results and give a lot of effective algorithms.But mostly achiev- ed in the PC,and can not meet the requirement of convenient, real-time and low cost .Under this situation,this paper put forward to achieve Adaboost face detection algorithm in the DSP environment, The main work of this paper is as follow:1. Face detection algorithms are mostly, after understanding the relevant algorithms and compareing their ,I chose Adaboost face detection algorithm and Its features is adaptable, small error rate. If the training samples is too good enough,cascade classifier based haar can make the error rate reduce to arbitrarily small, can be adapted to many complex environments.2. The paper use MATLAB to achieve Adaboost face detection algorithm,we can focus on the performance of algorithm,but not study how to achieve on their own.3.Using the 6000 DSP chips provide strong hardware suppor in the video image developed. DSP / BIOS real-time system programming divided into three tasks: as image processing, image analysis, image display,task can communication between each other.4.Opencv is open source computer vision library, a powerful tool for image processing, image engineers can significantly reduce product development time,The paper research face detec-detection of Opencv, I fully take into account the DSP structural features of the internal hard-ware and software programming in the transplant processing. Appropriate modifications and cutting used in the data structure and functions on Opencv can be reasonably carried out in the DSP System. |
PDF Full Text Request