CAJ | 학술논문

传统图像平滑能量衰减分析方法，对于人体运动图像高频分量抑制效果较差，且模糊图像边缘，运算效率较低。提出了一种依据替身运动DR算法的人体运动图像动态视点平滑能量衰减过程仿真方法，采用高斯滤波器对人体运动图像进行平滑处理，过滤人体运动图像的噪声，完整保留人体运动图像边缘信息以及细节，用替身描述人体在场景中的运动特征，替身在场景中运动并同其它人体的替身进行交互，采用局部化的人体模型提高人体运动场景的逼真度，通过人体运动图像的DR预测算法，完成人体运动图像动态视点平滑能量衰减过程分析。实验结果说明，所提方法能够确保人体运动图像边缘锐化，有效降低平滑能量、锐化边缘。
전통도상평활능량쇠감분석방법，대우인체운동도상고빈분량억제효과교차，차모호도상변연，운산효솔교저。제출료일충의거체신운동DR산법적인체운동도상동태시점평활능량쇠감과정방진방법，채용고사려파기대인체운동도상진행평활처리，과려인체운동도상적조성，완정보류인체운동도상변연신식이급세절，용체신묘술인체재장경중적운동특정，체신재장경중운동병동기타인체적체신진행교호，채용국부화적인체모형제고인체운동장경적핍진도，통과인체운동도상적DR예측산법，완성인체운동도상동태시점평활능량쇠감과정분석。실험결과설명，소제방법능구학보인체운동도상변연예화，유효강저평활능량、예화변연。
Traditional image smoothing energy attenuation analysis method, for the human body movement images high frequency component of the inhibition effect is poorer, and the fuzzy image edge, the computation efficiency is low. Put forward a kind of human movement image based on the double movement of DR algorithm dynamic viewpoint smooth energy attenuation process simulation method, the gaussian filter is adopted to human motion image smoothing, filtering the noise of the human body movement images, intact human body image edge information and details, human motion in the scene is described by using double, double motion in the scene and interact with other human body double, the localization of the human body model to improve the fidelity of human movement scene, by DR prediction algorithm of the human body movement images, complete human motion image dynamic viewpoint smooth energy attenuation process analysis. The experimental results indicate that the proposed method can ensure that human body image edge sharpening, effectively reduce the smooth energy, sharpen edges.