maturk

[Full images datamanager] fix image indexing with cv.undistort and implement eval indices logic

18

### Two things: #### 1) fix new undistorted image shape not being the same as input from my understanding in `newK, roi = cv2.getOptimalNewCameraMatrix(K, distortion_params, (image.shape[1], image.shape[0]), 0)` the roi...

Add more attributes to SplatFacto

3

For custom models, we need to access some of the intermediate outputs. Lets make them attributes so its easier to call `outputs = super(SplatFacto).get_outputs(camera)` and still have access to them.

Converting disparities to depths

1

Hi I am experimenting with using DepthAnything for depth supervision (not inverse depth == disparity). I am wondering if you have any advice in converting disparity estimates to relative depths,...

Inference on custom synthetic renders failing

1

Hi @zubair-irshad, I am trying to run some evaluations on my own dataset of renders of Synthetic ShapeNet models (same models you trained on); but, I am failing to run...

Made some improvements to the docs from my previous version. Mainly for clarity and reference.

[WIP] Unified depth rasterizer

6

Testing out unified rgb and depth rasterization in a single forward pass. Collaborating with @nikmo33 ! Currently these are the speeds on my machine (4090 GPU): Forward pass: - Unified...

Note, this PR does not have to be merged. It can stay as a PR or if people find it useful, it can be merged. As mentioned in #323, here...

gsplat metrics

1

Hey @jkulhanek, I am wondering if you are interested in adding the results of gsplat into your baseline comparisons as well? Currently, from my findings, some features in gsplat definitely...