Viet-Anh on Software Logo

What is: CodeSLAM?

SourceCodeSLAM - Learning a Compact, Optimisable Representation for Dense Visual SLAM
Year2000
Data SourceCC BY-SA - https://paperswithcode.com

CodeSLAM represents the 3D geometry of a scene using the latent space of a variational autoencoder. The depth thus becomes a function of the RGB image and the unknown code, D=Gθ(I,c)D = G_\theta(I,c). During training time, the weights of the network GθG_\theta are learnt by training the generator and encoder using a standard autoencoding task. At test time the code cc and the pose of the images is found by optimizing the reprojection error over multiple images.