obtain future art prediction (linear trend only, no VAR) by following the steps:
-
(1) use Classifier.py on the tinyImageNet dataset to train auxiliary classifier model.
-
(2) use WikiArt_crop.py then WikiArt_select_20_movements.py to produce art dataset. Make two versions: 64x64 and 128x128.
-
(3) use AE_perceptual with style and/or content loss to obtain the latent codes. Use the 64x64 version of the art dataset for this task.
-
(4) use CGAN_128.py to train the conditional GAN using the aforementioned images and codes.
-
(5) use Generate_future to make future art.
Post-Minimalism:
- https://www.wikiart.org/en/gego/untitled-1980
- https://www.wikiart.org/en/christopher-wilmarth/the-whole-soul-summed-up-1979
- https://www.wikiart.org/en/james-lee-byars/dress-for-five-persons-1969
- https://www.wikiart.org/en/takamatsu-jiro/jiro-takamatsu-1983
New Casualism: