LLaVA-3D could perform both 2D and 3D vision-language tasks. The left block (b) shows that compared with previous 3D LMMs, our LLaVA-3D achieves state-of-the-art performance across a wide range of 3D ...
Abstract: Three-dimensional coronary magnetic res- onance angiography (CMRA) requires reconstruction algorithms that can significantly suppress the artifacts encountered in heavily undersampled ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results