Abstract: Existing end-to-end image coding for machines (ICM) methods generally use joint training strategies to promote the compression efficiency for machine vision without considering the influence ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results