Synthetic Whiteimages for the Lytro Illum Light Field Camera with Ground Truth Microlens Center Coordinates

Citation Author(s):
Maximilian
Schambach
Karlsruhe Institute of Technology
Fernando
Puente León
Karlsruhe Institute of Technology
Submitted by:
Maximilian Schambach
Last updated:
Tue, 05/17/2022 - 22:21
DOI:
10.21227/msck-x083
Data Format:
Link to Paper:
Links:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

We provide a dataset with synthetic white images for the Lytro Illum light field camera with precisely known microlens center coordinates.

The dataset consists of white images taken at different zoom settings as well as different microlens array offset and rotation.

The white images have been raytraced using a thin lens-based camera model. The synthesized white images incorporate natural as well as mechanical vignetting effects.

For every microlens, a precise center coordinate is provided from which the orthogonal and perspective projections onto the camera sensor can be calculated.

For details on the camera model as well as usage of the dataset, please consult our main paper as well as the linked source code repositories.

Instructions: 

When using our dataset, please cite our corresponding paper:

M. Schambach and F. Puente León:

"Microlens Array Grid Estimation, Light Field Decoding, and Calibration,"

in IEEE Transactions on Computational Imaging, vol. 6, pp. 591-603, 2020

doi: 10.1109/TCI.2020.2964257.

 

The white images are provided as 16-bit PNG files and each white image is uniquely identified by a 8-digit hex code. For every white image, a .csv file is provided containing the microlens center position (in camera coordinates, in meters) where z=0 corresponds to the main lens plane. The sensor position (which is equal to the image distance of the camera configuration) can be calculated from the metadata which is provided in the .metadata file of each white image which is correpsonds to the JSON format. The metadata file provides:

  • the sensor size in px

  • the pixel pitch in m

  • the focus distance in m

  • the main lens radius in m

  • the microlens radius in m

  • the camera's f-number

  • the microlens array rotation angle in rad

  • the amount of grid noise

 

Please have a look at the provided GitLab repository (https://gitlab.com/iiit-public/papers/ml-grid-estimation-lf-decoding-and...) for examples of usage.

Dataset Files

LOGIN TO ACCESS DATASET FILES
Open Access dataset files are accessible to all logged in  users. Don't have a login?  Create a free IEEE account.  IEEE Membership is not required.