Fisheye Data Set

Andrea Eichenseer and André Kaup

A Data Set Providing
Synthetic and Real-World Fisheye Video Sequences

published in:

Proc. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 1541-1545, Shanghai, China, March 2016

Please note:

If you make use of any of our sequences, please do not forget to cite the paper
A Data Set Providing Synthetic and Real-World Fisheye Video Sequences, presented at ICASSP 2016.

The publication can be accessed via IEEE Xplore (abstract, pdf).

Copyright 2016 IEEE. Published in the IEEE 2016 International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2016), scheduled for 20-15 March 2016 in Shanghai, China. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works, must be obtained from the IEEE. Contact: Manager, Copyrights and Permissions / IEEE Service Center / 445 Hoes Lane / P.O. Box 1331 / Piscataway, NJ 08855-1331, USA. Telephone: + Intl. 908-562-3966.


This data set aims at providing researchers with video sequences for developing and testing algorithms designed for fisheye data.
Both synthetic sequences generated in Blender as well as real-world sequences captured with an actual fisheye camera are provided.

The field of view is 185° for all sequences; the focal length is 1.8 mm.
Further details on both setup and settings as well as on the sequence characteristics may be found in the corresponding ICASSP 2016 publication (posted above).

Information on licenses & credits may be found at the end of this page.

Synthetic sequences

Currently, only equisolid fisheye sequences are provided in PNG format (8 bit, RGBA). In the future, the synthetic sequences might be extended by equidistant (ED) fisheye versions.
To save storage space, the synthetic sequences have been compressed in a lossless manner using OptiPNG.

Click on the preview image for the original resolution (1088 x 1088). Click on the name below the preview image to download the sequence.

Cubes with Checkerboard Patterns

(same camera motion for all sequences)


use CheckercubeC
Download: CheckercubeA CheckercubeB CheckercubeC CheckercubeD

Cubes with Patterns and Text

(same camera motion for all sequencs)

Download: Rays Stars Lorem Alphabet Gradient

Image Cubes

(same camera motion for all sequences)

Download: Cards Clips Coins Fence Flowers Pencils

Automotive and Surveillance

(static camera)

Download: CarsA CarsB PoolNightB HallwayC LivingroomB Room

Surveillance “in Motion”

(moving camera including pure rotation and translation)

Download: Street HallwayA HallwayD HallwayB LivingroomA LivingroomC

Miscellaneous Scenes

(local, global, and mixed motion)

Download: PillarsA PillarsB PillarsC PoolA PoolB PoolNightA

Real-world sequences

Demosaicked 8-bit RGB sequences in PNG format are provided along with their corresponding 12-bit Bayer GR (RAW) versions in TIFF format.

Click on the preview image for the original resolution (1150 x 1086). Click on the name below the preview image to download the sequence.

Testchart and Text

(horizontal, vertical, shaky motion)

Download: TestchartA | RAW TestchartB | RAW TestchartC | RAW AlfaA | RAW AlfaB | RAW AlfaC | RAW

Book Shelves

(zoom, horizontal motion, horizontal pan, shaky motion)

Download: LibraryA | RAW LibraryB | RAW LibraryC | RAW LibraryD | RAW LibraryE | RAW


(moving camera)

Download: DriveA | RAW DriveB | RAW DriveC | RAW DriveD | RAW DriveE | RAW

Automotive and Surveillance

(static camera)

Download: CarparkA | RAW CarparkB | RAW CarparkC | RAW


(static camera)

Download: ElevatorA | RAW ElevatorB | RAW ElevatorC | RAW ElevatorD | RAW LabRoomA | RAW LabRoomB | RAW

Miscellaneous Scenes

(global motion)


(size of one square:
115 mm x 115 mm)

Download: ClutterA | RAW ClutterB | RAW LectureA | RAW LectureB | RAW Calibration

Licenses & Credits

The data set is publically, freely, and exclusively available for research purposes.
A compilation of the image files and blender models used for the synthetic sequences along with their respective licensing information is provided in this document: License Information


A huge thanks goes to the people who were involved in the creation of this data set. We would like to thank:

Michael Ortner for compiling the blender scenes and providing preview sequences
Andreas Luder for taking care of all things camera-related and capturing the real-world sequences
Michel Bätz and Martin Pöllot for helping out during capturing the indoor and outdoor scenes
Mira Hundhausen for helping evaluate and organize the thousands of images

This work was supported by the German Research Foundation (DFG) as Research Area A.3 of the Research Training Group GRK-1773 Heterogeneous Image Systems.


Andrea Eichenseer