Abstract
The existing RGB-D scene recognition approaches typically employ two separate and modality-specific networks to learn effective RGB and Depth represen......
小提示：本篇文献需要登录阅读全文，点击跳转登录
Full Text Link
