bob.ip.stereo 1.0.1

Stereo matching and image reprojection

Bob IP Stereo

This package is part of the signal-processing and machine learning toolbox Bob.

This package implements stereo mapping and projection functionalities and provides them as bob.io.stream.StreamFilter to easiily integrate them in processing pipelines.

Installation

Complete bob’s installation instructions. Then, to install this package, run:

$ conda install bob.ip.stereo

Example

Here is an example of how to use the package to load data recorded with 3 different cameras, build a depth map using 2 of them and use it to project the third camera’s data.

# Import Stream and StreamFile classes
# Generic processing filters are available through the Stream class
from bob.io.stream import StreamFile, Stream

# Import bob.ip.stereo: this updates the Stream class with the stereo filters
# It also upates the StreamFile class to allow to set camera configuration information.

f = StreamFile(
   resource_path("test/data/input_example.h5", "bob.io.stream"),
   resource_path("config/idiap_face_streams.json", "bob.io.stream"),
)
f.set_camera_configs(resource_path("config/idiap_face_calibration.json"))

# stream for stereo and projection tests
color = Stream("color", f)
nir_left = Stream("nir_left_stereo", f).adjust(color)
nir_right = Stream("nir_right_stereo", f).adjust(color)

# reproject operations
map_3d = nir_left.stereo(nir_right)
depth = map_3d.select(channel=2).colormap(colormap="jet")
rep_color = color.reproject(nir_left, nir_right, map_3d)

# Use the stereo data:
rep_color[0:2]  # ...

Contact

For questions or reporting issues to this software package, contact our development mailing list.