{"id":8931,"date":"2019-07-02T16:44:39","date_gmt":"2019-07-02T23:44:39","guid":{"rendered":"http:\/\/jnack.com\/blog\/?p=8931"},"modified":"2019-07-02T16:44:39","modified_gmt":"2019-07-02T23:44:39","slug":"ganpaint-promises-to-hallucinate-details-for-your-photos","status":"publish","type":"post","link":"http:\/\/jnack.com\/blog\/2019\/07\/02\/ganpaint-promises-to-hallucinate-details-for-your-photos\/","title":{"rendered":"GANpaint promises to hallucinate details for your photos"},"content":{"rendered":"<p>\u201cPaint using neurons instead of pixels,\u201d promises\u00a0<a href=\"http:\/\/gandissect.csail.mit.edu\/\">GAN Dissection<\/a>\u00a0a framework to \u201clet you explore what a GAN (generative adversarial networks) has learned by examining and manipulating its internal neurons.\u201d Check out how it can invent details like trees &amp; doorways based on the target image:<\/p>\n<p><iframe loading=\"lazy\" title=\"GAN Dissection: Visualizing and Understanding Generative Adversarial Networks\" width=\"604\" height=\"340\" src=\"https:\/\/www.youtube.com\/embed\/yVCgUYe4JTM?start=137&#038;feature=oembed\" frameborder=\"0\" allow=\"accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture\" allowfullscreen><\/iframe><\/p>\n<p>I\u2019m reminded of the O.G. PatchMatch demo from 10 (!) years ago that led us to put Content-Aware Fill (itself based on a subset of that work) into Photoshop:<\/p>\n<p><iframe loading=\"lazy\" title=\"PatchMatch: a randomized correspondence algorithm for structural imag.. (SIGGRAPH 2009 Presentation)\" width=\"604\" height=\"453\" src=\"https:\/\/www.youtube.com\/embed\/o1G9S-MvFos?feature=oembed\" frameborder=\"0\" allow=\"accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture\" allowfullscreen><\/iframe><\/p>\n<p>[<a href=\"https:\/\/youtu.be\/yVCgUYe4JTM?t=137\">YouTube 1<\/a> &amp; <a href=\"https:\/\/www.youtube.com\/watch?v=o1G9S-MvFos\">2<\/a>] [Via Product Hunt]<\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u201cPaint using neurons instead of pixels,\u201d promises\u00a0GAN Dissection\u00a0a framework to \u201clet you explore what a GAN (generative adversarial networks) has learned by examining and manipulating its internal neurons.\u201d Check out how it can invent details like trees &amp; doorways based on the target image: I\u2019m reminded of the O.G. PatchMatch demo from 10 (!) years [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[],"tags":[],"_links":{"self":[{"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/posts\/8931"}],"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=8931"}],"version-history":[{"count":1,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/posts\/8931\/revisions"}],"predecessor-version":[{"id":8932,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/posts\/8931\/revisions\/8932"}],"wp:attachment":[{"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/media?parent=8931"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/categories?post=8931"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/tags?post=8931"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}