Demosaicing Algorithm for Sony IMX250MYR

版本 1.1.0 (27.4 MB) 作者: Pierre-Jean Lapray

Demosaicing algorithm to demosaic colour polarization filter array images (for ex. from SONY IMX250MYR sensor)

https://github.com/pjlapray/LMMSE-Demosaicing-for-Colour-Polarization-Filter-Array

关注

0.0

(0)

56.0 次下载

更新时间 2024/3/9

在 GitHub 上查看许可证

Demosaicing-CPFA-LMMSE

This code allow you to demosaic a colour and polarimetric image from a filter array sensor with the Linear Minimum Mean Square Error demosaicing algorithm. The arrangement is the QuadBayer polarization pattern from the SONY IMX250MYR sensor. The monochrome version of this code is at : https://fr.mathworks.com/matlabcentral/fileexchange/131758-demosaicing-algorithm-for-sony-imx250-mzr/?s_tid=LandingPageTabfx. This code can also be applied to SONY IMX264MYR or IMX253MYR. The code will be readapted for any square arrangement in a future release.

The LMMSE demosaicing algorithm is a learning-based technique. The initial training (matrix 'D_matrix.mat' in Data folder) is done with the data used in the publication (training with 12 images from the Wen et al. database). If using the code, please cite these publications:

1- Dumoulin R., Lapray P.-J., Thomas J.-B., Farup I., Impact of training data on LMMSE demosaicing for Colour-Polarization Filter Array, 16th International Conference on Signal-Image Technology & Internet-Based Systems (SITIS), 2022, Dijon, France.

2- Spote A., Lapray P.-J., Thomas J.-B., Farup I., Joint demosaicing of colour and polarisation from filter arrays, In 29th Color and Imaging Conference Final Program and Proceedings 2021, Society for Imaging Science and Technology, 2021.

Two scripts are provided as demonstration:

The script to use for demosaicing with the furbished pre-trained matrix D is "Script_LMMSE_Demosaicing".
If you may want to retrain the demosaicing matrix with your own specific data, use the Script named "Script_LMMSE_retraining.m". Please use the same image structure as in "Data/Dataset". It is recommanded to train with sufficient data (at least 12 images of resolution 1456 × 1088 pixels, see our conference paper from 2022 for more information).

The "Data" folder contains a mosaiced image, a matrix used for demosaicing, along woth a dataset in case of retraining matrix D with the script "Script_LMMSE_retraining.m".

The "Function" folder contains Matlab functions needed for retraining.

引用格式

Pierre-Jean Lapray (2025). Demosaicing Algorithm for Sony IMX250MYR (https://github.com/pjlapray/LMMSE-Demosaicing-for-Colour-Polarization-Filter-Array/releases/tag/1.1.0), GitHub. 检索时间: 2025/9/6.

Dumoulin, Ronan, et al. “Impact of Training Data on LMMSE Demosaicing for Colour-Polarization Filter Array.” 2022 16th International Conference on Signal-Image Technology &Amp\Mathsemicolon Internet-Based Systems (SITIS), IEEE, 2022, doi:10.1109/sitis57111.2022.00031.

查看更多格式

Spote, Alexandra, et al. “Joint Demosaicing of Colour and Polarisation from Filter Arrays.” Color and Imaging Conference, vol. 29, no. 1, Society for Imaging Science & Technology, Nov. 2021, pp. 288–93, doi:10.2352/issn.2169-2629.2021.29.288.

查看更多格式

MATLAB 版本兼容性

创建方式 R2023a

兼容任何版本

平台兼容性

Windows macOS Linux

标签添加标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Function

版本	已发布	发行说明
1.1.0	2024/3/9	See release notes for this release on GitHub: https://github.com/pjlapray/LMMSE-Demosaicing-for-Colour-Polarization-Filter-Array/releases/tag/1.1.0	下载
1.0.0	2024/3/9	See release notes for this release on GitHub: https://github.com/pjlapray/LMMSE-Demosaicing-for-Colour-Polarization-Filter-Array/releases/tag/1.0.0	下载

要查看或报告此来自 GitHub 的附加功能中的问题，请访问其 GitHub 仓库。

MLA	Dumoulin, Ronan, et al. “Impact of Training Data on LMMSE Demosaicing for Colour-Polarization Filter Array.” 2022 16th International Conference on Signal-Image Technology &Amp\Mathsemicolon Internet-Based Systems (SITIS), IEEE, 2022, doi:10.1109/sitis57111.2022.00031.
APA	Dumoulin, R., Lapray, P.-J., Thomas, J.-B., & Farup, I. (2022). Impact of training data on LMMSE demosaicing for Colour-Polarization Filter Array. In 2022 16th International Conference on Signal-Image Technology &amp\mathsemicolon Internet-Based Systems (SITIS). IEEE. Retrieved from https://doi.org/10.1109%2Fsitis57111.2022.00031
BibTeX	@inproceedings{Dumoulin_2022, doi = {10.1109/sitis57111.2022.00031}, url = {https://doi.org/10.1109%2Fsitis57111.2022.00031}, year = 2022, month = {oct}, publisher = {{IEEE}}, author = {Ronan Dumoulin and Pierre-Jean Lapray and Jean-Baptiste Thomas and Ivar Farup}, title = {Impact of training data on {LMMSE} demosaicing for Colour-Polarization Filter Array}, booktitle = {2022 16th International Conference on Signal-Image Technology {\&}amp$\mathsemicolon$ Internet-Based Systems ({SITIS})} }

MLA	Spote, Alexandra, et al. “Joint Demosaicing of Colour and Polarisation from Filter Arrays.” Color and Imaging Conference, vol. 29, no. 1, Society for Imaging Science & Technology, Nov. 2021, pp. 288–93, doi:10.2352/issn.2169-2629.2021.29.288.
APA	Spote, A., Lapray, P.-J., Thomas, J.-B., & Farup, I. (2021). Joint demosaicing of colour and polarisation from filter arrays. Color and Imaging Conference, 29(1), 288–293. Society for Imaging Science & Technology. Retrieved from https://doi.org/10.2352%2Fissn.2169-2629.2021.29.288
BibTeX	@article{Spote_2021, doi = {10.2352/issn.2169-2629.2021.29.288}, url = {https://doi.org/10.2352%2Fissn.2169-2629.2021.29.288}, year = 2021, month = {nov}, publisher = {Society for Imaging Science {\&} Technology}, volume = {29}, number = {1}, pages = {288--293}, author = {Alexandra Spote and Pierre-Jean Lapray and Jean-Baptiste Thomas and Ivar Farup}, title = {Joint demosaicing of colour and polarisation from filter arrays}, journal = {Color and Imaging Conference} }