Recommended setup is through our Docker.
If instead you'd prefer a native install, go to: Setup_Instructions.
For every-time use, add the following lines (with paths adjusted) to your ~/.bashrc
LABELFUSION_SOURCE_DIR=/path/to/LabelFusion
DIRECTOR_INSTALL_DIR=/path/to/Director/install
source $LABELFUSION_SOURCE_DIR/setup_environment.sh
This is the quick version. If you'd prefer to go step-by-step manually, see Pipeline_Instructions.
First, cdlf && cd data/logs, then make a new directory for your data. In one terminal, run:
openni2-camera-lcm
In another, run:
lcm-logger
Your data will be saved in current directory as lcmlog-*.
First we will launch a log player with a slider, and a viewer. The terminal will prompt for a start and end time to trim the log, then save the outputs:
run_trim
Next, we prepare for object pose fitting, by running ElasticFusion and formatting the output:
run_prep
Next, launch the object alignment tool. See Pipeline_Instructions for how to use the alignment tool:
run_alignment_tool
After the alignment outputs have been saved, we can create the labeled data:
run_create_data
Navigate to /SegNet/MovingCamera/
Copy all the data you want to use (created by run_create_data from different datasets) into ./train
Use a different subdirectory inside /train/ for each log, i.e.:
/train/log-1 /train/log-2
Then resize all of the training images to a better size for training:
python resize_all_images.py
Finally, create the description of image-label pairs needed as SegNet input:
python create_traiing_set_list.py
To train SegNet:
cd / ./SegNet/caffe-segnet/build/tools/caffe train -gpu 0 -solver /SegNet/Models/moving_camera_solver.prototxt