I discussed with my supervisors and all of them agreed that there is undoubtedly a mistake in part of the measurements of a recently published paper. The idea and the novelty of this paper is correct. However, part of the measurements is incorrect. A figure and a part of the comparison table should be modified.
The authors tried to compare the new technique with the old one. The false measured results are shifted up with the same amount of increase (new and old) so this mistake will not affect on the final conclusion. Actually, the mistake does not affect the abstract or the conclusion of the paper.
I think IEEE transactions do not accept corrections, right? So, this is the reason why authors can not correct it. Should I communicate with the editor or authors, or just ignore it?