{"id":14731,"date":"2021-05-07T11:13:44","date_gmt":"2021-05-07T18:13:44","guid":{"rendered":"http:\/\/jnack.com\/blog\/?p=14731"},"modified":"2021-05-07T11:15:39","modified_gmt":"2021-05-07T18:15:39","slug":"interesting-interactive-mash-ups-powered-by-ai","status":"publish","type":"post","link":"http:\/\/jnack.com\/blog\/2021\/05\/07\/interesting-interactive-mash-ups-powered-by-ai\/","title":{"rendered":"Interesting, interactive mash-ups powered by AI"},"content":{"rendered":"\n<p>Check out how StyleMapGAN (<a href=\"https:\/\/www.youtube.com\/redirect?event=video_description&amp;redir_token=QUFFLUhqbWlCN210T0h6QkFmQzJDcl9vQk9PV3hFTmViZ3xBQ3Jtc0ttWnM2aHp4YjFKSGYybFRTLURDV2VLNlVEbDVrZ2NxNFhiZmN3cW9SbWhGeTQ0TjIyWVVEbl8zMXFWYXVsZEpick1jMUtmb1dnNVMxajgtcmVHQlZtYklPdGE1Vmdxb1hNZzdKT0M4el9oaGpwZFdJUQ&amp;q=https%3A%2F%2Farxiv.org%2Fabs%2F2104.14754\">paper<\/a>, <a href=\"https:\/\/arxiv.org\/pdf\/2104.14754.pdf\">PDF<\/a>, <a href=\"https:\/\/github.com\/naver-ai\/StyleMapGAN\">code<\/a>) enables combinations of human &amp; animal faces, vehicles, buildings, and more. Unlike simple copy-paste-blend, this technique permits interactive morphing between source &amp; target pixels:<\/p>\n\n\n\n<figure class=\"wp-block-embed is-type-video is-provider-youtube wp-block-embed-youtube wp-embed-aspect-16-9 wp-has-aspect-ratio\"><div class=\"wp-block-embed__wrapper\">\n<iframe loading=\"lazy\" title=\"StyleMapGAN: Exploiting Spatial Dimensions of Latent in GAN for Real-time Image Editing\" width=\"604\" height=\"340\" src=\"https:\/\/www.youtube.com\/embed\/qCapNyRA_Ng?feature=oembed\" frameborder=\"0\" allow=\"accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture\" allowfullscreen><\/iframe>\n<\/div><\/figure>\n\n\n\n<figure class=\"wp-block-image size-large\"><a href=\"http:\/\/jnack.com\/blog\/wp-content\/uploads\/2021\/05\/StyleMapGAN.jpeg\"><img decoding=\"async\" loading=\"lazy\" width=\"1024\" height=\"522\" src=\"http:\/\/jnack.com\/blog\/wp-content\/uploads\/2021\/05\/StyleMapGAN-1024x522.jpeg\" alt=\"\" class=\"wp-image-14744\" srcset=\"http:\/\/jnack.com\/blog\/wp-content\/uploads\/2021\/05\/StyleMapGAN-1024x522.jpeg 1024w, http:\/\/jnack.com\/blog\/wp-content\/uploads\/2021\/05\/StyleMapGAN-300x153.jpeg 300w, http:\/\/jnack.com\/blog\/wp-content\/uploads\/2021\/05\/StyleMapGAN-768x392.jpeg 768w, http:\/\/jnack.com\/blog\/wp-content\/uploads\/2021\/05\/StyleMapGAN-1536x783.jpeg 1536w, http:\/\/jnack.com\/blog\/wp-content\/uploads\/2021\/05\/StyleMapGAN.jpeg 1920w\" sizes=\"(max-width: 1024px) 100vw, 1024px\" \/><\/a><\/figure>\n\n\n\n<p>From the authors, a bit about what&#8217;s going on here:<\/p>\n\n\n\n<blockquote class=\"wp-block-quote\"><p>Generative adversarial networks (GANs) synthesize realistic images from random latent vectors. Although manipulating the latent vectors controls the synthesized outputs, editing real images with GANs suffers from i) time-consuming optimization for projecting real images to the latent vectors, ii) or inaccurate embedding through an encoder. We propose StyleMapGAN: the intermediate latent space has spatial dimensions, and a spatially variant modulation replaces AdaIN. It makes the embedding through an encoder more accurate than existing optimization-based methods while maintaining the properties of GANs. Experimental results demonstrate that our method significantly outperforms state-of-the-art models in various image manipulation tasks such as local editing and image interpolation. Last but not least, conventional editing methods on GANs are still valid on our StyleMapGAN. Source code is available at <a rel=\"noreferrer noopener\" href=\"https:\/\/www.youtube.com\/redirect?event=video_description&amp;redir_token=QUFFLUhqbFZCVjJfSWx3VVBVeDRXdG5OUGZrM2JuaE5Id3xBQ3Jtc0trMUN4c0MzRlRqVElKcXNIbzJDZW95bnB5bTNSQ3pnV2N1SjNhUzNQNWpYckJIRnZUNmdDLW1iR1BISEt4MnhXejEtN2NtbnBxMmJ1QmpqRzBYMjA2dzYwZFBzWk5KdlVieEJsam1VOS12ejdSUTdwOA&amp;q=https%3A%2F%2Fgithub.com%2Fnaver-ai%2FStyleMapGAN\" target=\"_blank\">https:\/\/github.com\/naver-ai\/StyleMapGAN<\/a>\u200b.<\/p><\/blockquote>\n\n\n\n<p><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Check out how StyleMapGAN (paper, PDF, code) enables combinations of human &amp; animal faces, vehicles, buildings, and more. Unlike simple copy-paste-blend, this technique permits interactive morphing between source &amp; target pixels: From the authors, a bit about what&#8217;s going on here: Generative adversarial networks (GANs) synthesize realistic images from random latent vectors. Although manipulating the [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[66,3],"tags":[],"_links":{"self":[{"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/posts\/14731"}],"collection":[{"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/comments?post=14731"}],"version-history":[{"count":4,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/posts\/14731\/revisions"}],"predecessor-version":[{"id":14747,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/posts\/14731\/revisions\/14747"}],"wp:attachment":[{"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/media?parent=14731"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/categories?post=14731"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/tags?post=14731"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}