Closed thucz closed 1 year ago
Yes, novel view synthesis is supported. I think you first need to convert your colmap data in our format and then train the model. After training, you can render RGB images from any viewpoints. We have some novel view synthesis results on the paper Figure 18.
Recently, I submitted an article and the reviewer asked me to compare my method with MonoSDF for the novel view synthesis task. How can I run MonoSDF for the novel view synthesis task with the poses of COLMAP?