Benchmarking Deepart Detection

Figure 1: Examples of the established deepart detection database (DDDB). The examples of LAION-5B Schuhmann et al. (2022) are conventional artworks (conarts), and the rest examples (i.e., StableDiff Rombach et al. (2021),DALL-E 2 Ramesh et al. (2022),Imagen Saharia et al. (2022),Midjourney Holz (2022), and Parti Yu et al. (2022)) are deepfake artworks (deeparts) produced by generative models. Our data and code will be released. Deepfake technologies have been blurring the boundaries between the real and unreal, likely resulting in malicious events. By leveraging newly emerged deepfake technologies, deepfake researchers have been making a great upending to create deepfake artworks (deeparts), which are further closing the gap between reality and fantasy. This database enables us to explore once-for-all deepart detection and continual deepart detection. For the two new problems, we suggest four benchmark evaluations and four families of solutions on the constructed DDDB. The comprehensive study demonstrates the effectiveness of the proposed solutions on the established benchmark dataset, which is capable of paving a way to more interesting directions of deepart detection. The constructed benchmark dataset and the source code will be made publicly available. There has been a propensity to view deepfake technologies as destructive to the supposed boundaries between the real and unreal, leading to potentially detrimental effects. Despite this, deepfake researchers are continuing to make breakthroughs by wielding newly emerged deepfake technologies to create artworks, which are called deeparts throughout this paper. The new deepart techniques include Stable DiffusionRombach et al. (2021), DALL-E Ramesh et al. (2021; 2022), Imagen Saharia et al. (2022), Midjourney Holz (2022), and Parti Yu et al. (2022) As shown in Figure 1, compared to conventional deepfakes, deeparts have been making the boundary between reality and fantasy much more blurry.