depthy

Description

depthy enables depth map generation from light-fields and stereo image pairs.

Light-field Results

Central view
Depth map

source: HCI data

Stereo Results

Left Image	Ground Truth	Depthy (SGM)	OpenCV

source: Middlebury data

Installation

• via pip:

  1. install with pip3 install depthy
  2. type depthy -h to the command line once installation finished

• from source:

  1. install Python from https://www.python.org/
  2. download the source using git clone https://github.com/hahnec/depthy.git
  3. go to the root directory cd depthy
  4. load dependencies $ pip3 install -r requirements.txt
  5. install with python3 setup.py install
  6. if installation ran smoothly, enter depthy -h to the command line

Command Line Usage

From the root directory of your downloaded repo, you can run the tool on the provided test data by

depthy -s './examples/data/pens/'

on a UNIX system where the result is found at ./examples/data/. A windows equivalent of the above command is

depthy --src=".\\examples\\data\\pens\\"

To infer light-field depth, you can specify the method or select your image folder manually with

depthy --win --method='epi'

For stereo depth, you would use

depthy --win --method='stereo'

More information on optional arguments, can be found using the help parameter

depthy -h

API Usage

A documentation of the key functions is found at https://hahnec.github.io/depthy/. A reference integration using depthy's API can be found in PlenoptiCam.

Author

Christopher Hahne

Citation

@misc{depthy,
    title={Depthy},
    author={Christopher Hahne},
    year={2021},
    publisher = {GitHub},
    journal = {GitHub repository},
    howpublished = {\url{https://github.com/hahnec/depthy}}
}