CAJ | 학술논문

为了满足图像和视屏中人脸的位置，朝向和旋转都不是固定的特点，利用LVQ神经网络模型对图像中人脸朝向识别进行研究．实验采用Matlab工具箱进行LVQ神经网络设计，实现对人脸朝向的判断．实验结果显示，LVQ神经网络可以根据输入图像的二值信息，以较高准确率判断该图像的人脸朝向．
위료만족도상화시병중인검적위치，조향화선전도불시고정적특점，이용LVQ신경망락모형대도상중인검조향식별진행연구．실험채용Matlab공구상진행LVQ신경망락설계，실현대인검조향적판단．실험결과현시，LVQ신경망락가이근거수입도상적이치신식，이교고준학솔판단해도상적인검조향．
In order to meet the characteristics that the face position, orientation and the rotating are not fixed in the images and videoes, face orientation recognition is studied by using LVQ neural network model in the paper. The experiment using Matlab toolbox for the design, realization of LVQ neural network can identify face orientation, the experimental results show, according to the binary information of the input image, LVQ neural netwok can identify face orientation with higher accuracy.