Abstract: Recent advancements in generative models have significantly improved novel view synthesis (NVS) from multi-view data. However, existing methods depend on external multiview alignment ...
Abstract: This work investigates the problem of efficiently learning discriminative low-dimensional (LD) representations of multiclass image objects. We propose a generic end-to-end approach that ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback