search

siddiquesalman / flatnet

This is the official pytorch code repo for the ICCV 2019 paper Towards Photorealistic Reconstruction of Highly Multiplexed Lensless Images and the TPAMI 2020 paper FlatNet: Towards Photorealistic Scene Reconstruction from Lensless Measurements
BSD 3-Clause "New" or "Revised" License
31 stars 11 forks source link

About how to get the intermediate image #2

Closed lilubit closed 2 years ago

lilubit commented 2 years ago

Good job! It seems that the variable “Xout” in function “FlatNet” is the intermediate result. But it's 14256*256. Maybe some conversion to RGB space is needed? Looking forward to your reply！

siddiquesalman commented 2 years ago

Hi, if you're using a Batchx4x500x620 input to FlatNet, then Xout will be Batchx256x256x4. We have 4 channels because we are directly working in the Bayer space. To get RGB from this 4 channel Xout, you can either throw away one of the green channels or take the average of the two green channels.

lilubit commented 2 years ago

Got it! Thanks a lot!