摘要
针对目前视频对象分割算法中存在分割速度慢的问题,提出了一种基于LDOF光流算法的视频对象分割方法。首先,利用LDOF((Large Displace Optical Flow)计算视频帧中t与t+1时刻像素点的光流;然后,计算N个邻近像素的方向导数,获得视频对象粗略轮廓信息;再利用高斯混合模型建立前景与背景模型,并运用扫描线算法获得到视频对象精确的轮廓;最后,采用Grab Cut算法对视频对象进行分割。实验结果表明,该算法能有效的提高了视频对象分割的速度。
According to the problems of slow segmentation in the video object segmentation,a video object segmentation method was proposed based on LDOF optical flow algorithm. Firstly,the pixel optical flow in times of t to t + 1 were calculated out by the method of LDOF( Large Displace Optical Flow,LDOF). Then,we calculate the directional derivative of N neighboring pixels to obtain the information of the video's rough outline. By using the Gaussian mixture model to build the foreground and background models and using the algorithm of scan line to calculate out the video object that has more precise contours. Finally,the video object is segmented by using the Grab Cut algorithm. The experimental results show that the algorithm is effective to improve the speed of video object segmentation.
引文
[1]张桂梅,王大雷.结合LPG&PCA的中智学图像分割[J].中国图象图形学报,2014,19(5):693-700.
[2]曾接贤,季康.基于多特征融合的多视点飞机目标识别算法[J].南昌航空大学学报:自然科学版,2016,30(2):8-15.
[3]储珺,施芒,符祥.基于光流的动态背景运动目标检测算法[J].南昌航空大学学报:自然科学版,2011,25(3):1-6.
[4]李宗民,公绪超,刘玉杰.多特征联合建模的视频对象分割技术研究[J].计算机学报,2013,36(11):2356-2363.
[5]祝世平,郭智超,高洁,等.基于Snake活动轮廓模型的视频跟踪分割方法[J].光电子·激光,2013,24(1):139-145.
[6]李小和,张太镒,周亚同,等.基于高斯混合模型的视频对象分割算法[J].西安交通大学学报,2006,40(6):724-728.
[7]Bai X,Wang J,Sapiro G.Dynamic color flow:a motion-adaptive color model for object segmentation in video[C]//European Conference on Computer Vision(ECCV),2010:617-630.
[8]Brox T,Malik J.Object segmentation by long term analysis of point trajectories[C]//In European Conference on Computer Vision(ECCV),2010:282-295.
[9]Lee Y J,Kim J,Grauman K.Key-segments for video object segmentation[C]//In IEEE International Conference on Computer Vision(ICCV),2011:1995-2002.
[10]T Brox,J Malik.Large displacement optical flow:descriptor matching in variational motion estimation[J].IEEE Transactions on pattern analysis and machine intellignce,2011,33(3):1-14.
[11]Ma T,Latecki L J.Maximum weight cliques with mutex constraints for video object segmentation[C]//In Conference on Computer Vision and Pattern Recognition(CVPR),2012:670-677.
[12]Zhang D,Javed O,Shah M.Video object segmentation through spatially accurate and temporally dense extraction of primary object regions[C]//In Conference on Computer Vision and Pattern Recognition(CVPR),2013:628-635.
[13]Rother C,Kolmogorov V,Blake A.“grabcut”:interactive foreground extraction using iterated graph cuts[J].ACM Transactions on Graphics,2004,23(3):309-314.