About keyframes times of use
lengfeng-bot opened this issue · 2 comments
First of all, thank you for your excellent work ! I ran your SLAM program before. The result printed out a list of the number of keyframes used, which I found to be used many times. So I would like to ask, for the part of Gaussian mapping, whether to only render keyframes and compute the loss of the rendered keyframes image and the ground truth. In this case, is the ground truth of non-keyframes wasted, and non-keyframe pictures are not actually rendered. what is the purpose and benefit of doing so?
I'm curious about this, because I think it's an important part and I hope you can answer my question.Thanks!
Hi. Thanks for your interest. Yes, we perform mapping optimization among keyframes. The main reason is to reduce computational cost. Using all frames may increase the final rendering quality but it will take more time for optimization.
Thank you very much for your answer, I will continue to read your code and keep learning