InfiniTAM
InfiniTAM copied to clipboard
Algomorph
A Framework for the Volumetric Integration of Depth Images (with some experimental code to handle dynamic scenes)
The Killing regularizer seems to grow out of proportion with every iteration, whether or not truncated voxels are included in the computation.
The method you employed to calculate the gradient of the killing term in 'SmoothingTermSharedRoutines.h' appears to differ from the approach described in the original paper. Could you please provide a...
The new tracker will stick to the "Engine" convention for now. We can tentatively call it DenseColorAndDepthRigidAlignmentEngine, but it might make sense to combine the different tracker variants in a...
Occasionally the user will need to: 1) Compare all values of two arrays after they are sorted for each array in a particular order (presort) 2) Compare all values of...
1. OStreamWrapper and IStreamWrapper need to have > operators defined for basic and primitive types, such as int, float, double, string, etc. 2. They should probably also be renamed to...
This can be used to address multiple issues in the code in the future, e.g. valid depth counting in image processing and the like. The engine should accept two lambdas:...
Investigate why this happens. It may very well be that the bug is in the original InfiniTAM code, but it also might be that the DepthIntegrationEngine doesn't integrate color properly...
The code is unnecessarily confusing due to the integer indexing of live volumes as "source" and "live" and then swapping the two indices. Use two pointers directly, and swap the...
- [x] Exponential learning_rate - [x] Linear learning_rate - [ ] Exponential warp_update_threshold - [ ] Linear warp_update_threshold
Pick two parameter sets, run the simulation with continuous mesh output and compile Hausdorff distance RMS from original Snoopy on each, plot this. See how error fluctuates. This may help...