A method of caption location and segmentation in news video | IEEE Conference Publication | IEEE Xplore