Fast Feature-Less Quaternion-based Particle Swarm Optimization for Object Pose Estimation From RGB-D Images
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
GraphCannySeg.cpp
GraphCannySeg.h
Makefile
README.md
cutil_math.h
handle_error.h
main.cu
main.h
myTimer.h

README.md

Fast Feature-Less Quaternion-based Particle Swarm Optimization for Object Pose Estimation From RGB-D Images

The code in this repository is related to the original paper: http://www.bmva.org/bmvc/2016/papers/paper113/ and includes some improvements.

This is the GPU implementation, which includes object segmentation from https://arxiv.org/abs/1605.03746


Hardware:

Requires a CUDA enabled GPU (with support for CUDA 7.0 or greater)

Sofware:

  • OpenCV
  • gcc with support for c++11

Compile:

make

Run (estimate the pose of the first 15 detected object clusters):

for N in {0..15}; do; ./main object_model rgb_image depth_image $N; done

Where rgb_image and depth_image are 640x480 png images.

For tuning the object segmentation parameters for different datasets, it could be useful to use the code from https://github.com/morpheus1820/graph-canny-segm