Classical Gas: 3000 Years of Art - a reconstruction
This project dissembles a copy of the original Classical Gas, 3000 Years of Art video and identifies all the artworks used.
Copies in better condition, and in larger formats, are found on the web and saved to the project.
Help would be very valuable in this project. Therer are over 2,000 images in the video that need to be identified. Many are smaller subsections of an artwork rather than the entire work. If a large enough copy cannot be found, I will attempt enlargement using one of the current deep-learning super resolution techniques.
To date, I have identified and downloaded over 100 artworks.
The Google search engine for images, helps a lot. That service can upload an image of a work of art, or a subsection of a work of art, and attempt to find the same or similar images on the web. Sometimes the search is quick and exact, other times, way off. One image was used as a cover for a recent book. The engine could not get away from the book reference, subsequently, art images were not searched as thoroughly as was possible. Sometimes, the greenish color of the originals is correctable and results in a success. Other times color correction combined with sharpening will yield a success. Once, I simply found the publishing company of a book and sent an email asking if anyone knew who had painted the picture on the book cover. I got a working response.
There will be no single approach to get all the images identified.
I am still looking for help with this project. Without help, the project can easily take three months of full time work.
If I download 100 artworks each day, twenty days will have passed before I complete the identification process. Then, the rebuilding can take place.
Is the goal worth the time and effort?
I don't know. I may keep my mind open for something more fundamental and longlasting to do.
Maybe, I'll go on vacation.
Who knows.
Copies in better condition, and in larger formats, are found on the web and saved to the project.
Help would be very valuable in this project. Therer are over 2,000 images in the video that need to be identified. Many are smaller subsections of an artwork rather than the entire work. If a large enough copy cannot be found, I will attempt enlargement using one of the current deep-learning super resolution techniques.
To date, I have identified and downloaded over 100 artworks.
The Google search engine for images, helps a lot. That service can upload an image of a work of art, or a subsection of a work of art, and attempt to find the same or similar images on the web. Sometimes the search is quick and exact, other times, way off. One image was used as a cover for a recent book. The engine could not get away from the book reference, subsequently, art images were not searched as thoroughly as was possible. Sometimes, the greenish color of the originals is correctable and results in a success. Other times color correction combined with sharpening will yield a success. Once, I simply found the publishing company of a book and sent an email asking if anyone knew who had painted the picture on the book cover. I got a working response.
There will be no single approach to get all the images identified.
I am still looking for help with this project. Without help, the project can easily take three months of full time work.
If I download 100 artworks each day, twenty days will have passed before I complete the identification process. Then, the rebuilding can take place.
Is the goal worth the time and effort?
I don't know. I may keep my mind open for something more fundamental and longlasting to do.
Maybe, I'll go on vacation.
Who knows.
Comments
Post a Comment