Tencent CNN in MRF: Video Object Segmentation SpatioTemporal MRF
Thispaperaddressestheproblemofvideoobjectsegmentation,wheretheinitialobjectmaskisgiveninthefirstframeofaninputvideo.WeproposeanovelspatiotemporalMarkovRandomField(MRF)modeldefinedoverpixelstohandlethisproblem.UnlikeconventionalMRFmodels,thespatialdependenciesamongpixelsinourmodela
下载地址
用户评论
