The Image Survival and Protection of Intangible Cultural Heritage in the New Media Environment

Abstract

As a product of agricultural civilization, intangible cultural heritage (ICH) has been in a bad situation in recent years. Modern video media, with the dual identity of art and media, is an effective way to preserve and disseminate ICH. First, a hybrid network composed of a Bi-directional Long Short-Term Memory (Bi-LSTM) network with attention structure and Neural Network is adopted to extract relevant knowledge. Then, the generative adversarial network (GAN) is optimized. Lastly, this model is tested. The test results reveal that in the dataset constructed here, when the resolution of the processed image is 48×64×48, it takes 0.4825s for the unimproved GAN to process the image, while the algorithm improved only needs 0.0391s to process the image, with a speedup of 12.2.