What if ViTs could do classical computer vision? Our #3DV2022 work shows with minor changes, a ViT performs similar computations to the 8-point algorithm.💡 This is a useful inductive bias for estimating camera pose! abs: arxiv.org/abs/2208.08988 ✍️ @jcjohnss and David Fouhey
4
38
195
0
46
Download Image
Proj: crockwell.github.io/rel_pose/ Github: github.com/crockwell/rel_… Key Takeaways 🧵
@_crockwell @CSProfKGD @jcjohnss How about evaluating it on really challenging problem? #IMC2022 is over, but kaggle allows for late submissions. Format: input: 2 images, output: fundamental matrix. kaggle.com/competitions/i…