search

DanielCoelho112 / synfeal

Synfeal: A Data-Driven Simulator for Camera Localization
9 stars 1 forks source link

Create neural network that receives both the RGB and depth images #49

Open DanielCoelho112 opened 2 years ago

DanielCoelho112 commented 2 years ago

Experiment early fusion and middle fusion (VMLoc: Variational Fusion For Learning-Based Multimodal Camera Localization)

miguelriemoliveira commented 2 years ago

If you want we can try to meet to try to do the 4 channel approach ... so I learn a bit more.

DanielCoelho112 commented 2 years ago

Sure, right now I am preparing the normalization of the rgb images, and also doing some final hyperparameter optimization of the depthnet. Next week if you are available we can do it.

miguelriemoliveira commented 2 years ago

OK, let me know next week and we arrange a meeting.

DanielCoelho112 commented 2 years ago

Hi @miguelriemoliveira,

the dataset loader is already ready for this issue. Let me know when you are available and then we can build the 4-channel version.

miguelriemoliveira commented 2 years ago

Hi @DanielCoelho112 ,

I am completely overwhelmed right now, and I think you too. Lets think about this in a few weeks.