CAJ | 학술논문

将人脑视觉注意机制应用于人脸图像分割与跟踪中,提出了一种基于视觉显著特征的人脸目标分割与跟踪算法.该方法由三步完成：首先通过模拟人脑视觉注意机制迅速而准确地利用颜色、结构、梯度和位置等信息建立人脸显著特征图.其次,基于建立的视觉显著特征图,对人脸图像视觉显著特征进行学习和聚类,最终能够快速而准确地确认和分割出图像中的人脸区域.该方法突破了传统的逐点搜索的限制,通过一个几何模型和眼图模型对图像中的人脸区域进行搜索,大大提高了人脸候选区域搜索标记的效率,减少了后续处理工作.最后,通过分割出的人脸区域得到一个有效的边界特征图,并融合人脸显著特征图对人脸进行跟踪.实验结果表明本论文所提出的基于视觉显著特征的人脸图像分割与跟踪方法能够较有效地分割出人脸.
장인뇌시각주의궤제응용우인검도상분할여근종중,제출료일충기우시각현저특정적인검목표분할여근종산법.해방법유삼보완성：수선통과모의인뇌시각주의궤제신속이준학지이용안색、결구、제도화위치등신식건립인검현저특정도.기차,기우건립적시각현저특정도,대인검도상시각현저특정진행학습화취류,최종능구쾌속이준학지학인화분할출도상중적인검구역.해방법돌파료전통적축점수색적한제,통과일개궤하모형화안도모형대도상중적인검구역진행수색,대대제고료인검후선구역수색표기적효솔,감소료후속처리공작.최후,통과분할출적인검구역득도일개유효적변계특정도,병융합인검현저특정도대인검진행근종.실험결과표명본논문소제출적기우시각현저특정적인검도상분할여근종방법능구교유효지분할출인검.
In this thesis,a new face segmentation and tracking algorithm based on visual salient features is proposed according to the human brain visual attention mechanism.The proposed method consists of three stages.The first stage is to simulate the human brain visual attention mechanism and establish a face saliency map according to color,structure,gradient and location information.The second stage is to identify the face region in the image and segment the face region quickly and accurately by learning and clustering the face salient features.This method improves the efficiency of the face candidate region searching,reduces the subsequent processing by building a geometric model and an eye-map,and outperforms the traditional point-by-point searching mechanism.The third stage is to get an effective boundary saliency map from the segmented face region and then track the face based on the segmented result.The experimental results show that the proposed visual salient feature based method is capable of segmenting the face area quite effectively.