Computer Vision News - May 2023

7 NeRF from Sparse and Noisy Poses it was initially, and I thought, how are we ever going to be able to get something reasonable? Now, we get something that looks realistic. ” Fabian adds: “ The most important thing is that rather than having to record for hours, you can just take a bunch of images or crawl the internet. That opens a lot of doors and makes it more accessible for everybody! ” As the first author of the work, Prune will present at CVPR 2023 in June . She worked on the codeduringher internshipat Google , where in the beginning, she was a newbie to the NeRF field and can remember being surprised at the results when the state-of- the-art methods were applied to sparse views. “ Iwas like, howcan it besobad? ” she recalls. “ I’mmost proud of the difference in outputs we can now get. That’s because I saw how