Dust and Scratch removal
As an effort to train and test a scratch and dust removal tool that we have been working on, we now require help from the community. At the moment we would like to request film scans that have been treated for scratch and dust removal, submitting the "before" and "after" images, or clean scans that we can use to simulate scratch and dust. This will help us testing the feature with actual film scans, so that it is able to deal with grain without thinking that it is dust.
To submit your images, please upload a zip to the following link: https://www.dropbox.com/request/Sq1921j4ZBDSEDBa40fn
These are the guidelines for submission:
- Upload a zip containing the images you want to submit. You can use your name, or instagram handle to name the file, so the name remains unique. (For example, _oscarbolanos_.zip). You can upload more than one zip, just make sure that they have different names as Dropbox overrides the data if it has the same name.
- Naming conventions. To submit a before/after pair, please add "-before" suffix and “-after" suffix to the source and corrected images. Optionally, you can include the film type and ISO in the file name (for example: "HP5@1600 1-before.jpeg" and "HP5@1600 1-after.jpeg"). For clean scan submissions, just mention the film type. They can be in any format (DNG, JPEG, TIFF), but please respect the file size restrictions (20MB).
- Images should be at least 320px on the shorter side. Although there is no technical limitation, please do not submit bigger than 20MB per image as my Dropbox will implode.
- Corrected images for dust and scratches should contain no color nor sharpness variations, meaning the only difference between the "before" and the "after" image is the elimination of dust and scratches. Images can include frames, but no water marks!
- Both Color and Black & White images are accepted.
- You can optionally include a README file with your contact information in case we have any questions.
By submitting an image you accept that it will be used for testing and training purposes. In case we would be interested in show casing the results using any of your submitted images, we would contact you separately to ask for permission.
Please do not worry about the quality of the images as the purpose is to evaluate the technology.