Lp-TV model for structure extraction with end-to-end contour learning | IEEE Conference Publication | IEEE Xplore