Version: 1.1.0

This repository contains the code corresponding to our CVPR 2020 paper: Tangent Images for Mitigating Spherical Distortion.

The linked paper is the arXiv version, which we have updated with additional experiments and which contains results that matches the code provided here.

This repository is designed to be used with PyTorch. This code requires the installation of my Spherical Distortion Package, which can be found here. Installation instructions are available in the linked repository.

You should be able to test if the installation was successful by running the example scripts in examples.

In the examples directory, we have provided some basic examples to help you get started using tangent images. These examples include rendering to and from tangent images as well some visualizations of tangent images for SIFT keypoint and Canny edge detection.

All experiments are included in the experiments folder. Each experiment subdirectory has a README file explaining how to setup and run each experiment. Where relevant, we have included the pre-trained models corresponding to our published results. Note that the semantic_segmentation directory contains both our standard semantic segmentation experiments as well as the network transfer experiments.

If you find this repository useful for your own work, please make sure to cite our paper:

@InProceedings{Eder_2020_CVPR,
    author = {Eder, Marc and Shvets, Mykhailo and Lim, John and Frahm, Jan-Michael},
    title = {Tangent Images for Mitigating Spherical Distortion},
    booktitle = {The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    month = {June},
    year = {2020}
}