Abstract: A key challenge in visible-infrared person re-identification (V-I ReID) is training a backbone model capable of effectively addressing the significant discrepancies across modalities.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results