Invitation to check out my FFT Image Restoration Project

[gabriel-filincowsky]

Hi @rechmbrs, @fpierfed,

I’ve recently developed an open-source tool that might align with your interests. It’s called FFT Pattern Suppression, and it's designed for automatic detection and removal of unwanted periodic patterns in images—like halftone dots and paper textures—using FFT (Fast Fourier Transform) techniques.

The tool offers real-time visualization and GPU acceleration via CuPy (CUDA) for performance, but also runs on CPU for broader compatibility. It could be useful for anyone in digital restoration, photography, or archival work.

I'd love to hear your thoughts or feedback, especially because this repo was inspired on your work. You can check it out here: https://github.com/gabriel-filincowsky/FFT-Pattern-Suppression

Best regards, Gabriel

[chain83]

Hi

Sounds good. A (near) real-time preview is something that can be very helpful when doing the finer adjustments and figuring out how it works, and something lacking with the filter+action approach .

The installation process is a bit involved, so I would recommend trying to package it into something a bit more approachable for the less tech-savvy users. I haven't found the time to test it myself (and I don't have an NVIDIA GPU at the moment).

On Mon, Sep 23, 2024 at 7:34 PM gabriel-filincowsky < @.***> wrote:

Hi @rechmbrs https://github.com/rechmbrs, @fpierfed https://github.com/fpierfed,

I’ve recently developed an open-source tool that might align with your interests. It’s called FFT Pattern Suppression, and it's designed for automatic detection and removal of unwanted periodic patterns in images—like halftone dots and paper textures—using FFT (Fast Fourier Transform) techniques.

The tool offers real-time visualization and GPU acceleration via CuPy (CUDA) for performance, but also runs on CPU for broader compatibility. It could be useful for anyone in digital restoration, photography, or archival work.

I'd love to hear your thoughts or feedback, especially because this repo was inspired on your work. You can check it out here: https://github.com/gabriel-filincowsky/FFT-Pattern-Suppression

Best regards, Gabriel

— Reply to this email directly, view it on GitHub https://github.com/rechmbrs/FtPattern/issues/6, or unsubscribe https://github.com/notifications/unsubscribe-auth/AZTRJ4ZYPHKQJY6MWPHLNCTZYBGJJAVCNFSM6AAAAABOWR75VOVHI2DSMVQWIX3LMV43ASLTON2WKOZSGU2DGMRVGEZTKNA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

-- Mvh. Jonas M. Rogne

[fpierfed]

Hi there,

Congrats for getting the tool out! It does look quite good. I tested it with a simple image on my Mac and it works great. I just created a PR for a trivial fix in the requirement files but that was it.

On 23 Sep 2024, at 19:34, gabriel-filincowsky @.***> wrote:

Hi @rechmbrs https://github.com/rechmbrs, @fpierfed https://github.com/fpierfed,

I’ve recently developed an open-source tool that might align with your interests. It’s called FFT Pattern Suppression, and it's designed for automatic detection and removal of unwanted periodic patterns in images—like halftone dots and paper textures—using FFT (Fast Fourier Transform) techniques.

The tool offers real-time visualization and GPU acceleration via CuPy (CUDA) for performance, but also runs on CPU for broader compatibility. It could be useful for anyone in digital restoration, photography, or archival work.

I'd love to hear your thoughts or feedback, especially because this repo was inspired on your work. You can check it out here: https://github.com/gabriel-filincowsky/FFT-Pattern-Suppression

Best regards, Gabriel

— Reply to this email directly, view it on GitHub https://github.com/rechmbrs/FtPattern/issues/6, or unsubscribe https://github.com/notifications/unsubscribe-auth/AADY6TGICG6CVJOFI77IOVLZYBGJJAVCNFSM6AAAAABOWR75VOVHI2DSMVQWIX3LMV43ASLTON2WKOZSGU2DGMRVGEZTKNA. You are receiving this because you were mentioned.

[gabriel-filincowsky]

@chain83 Thank you for the feedback; much appreciated.

[gabriel-filincowsky]

@fpierfed Good catch. Thank you for your contribution.