CORTEXA
← Browse

Hongyu Yang

1 paper indexed

arxivcs.CVcs.LG2026-07-24

Rethinking Multi-Branch and Cross-Backbone Fusion for Vehicle Re-Identification in the Foundation-Model Era

Yu Wang, Hongyu Yang

Multi-branch architectures and CNN-Transformer fusion have long been regarded as effective ways to improve vehicle re-identification (Re-ID) by combining complementary representations. In this work, we revisit this assumption in the foundation-model era through a comprehensive em…

View free PDFSource page