科技藝術書報討論
AIMS FELLOWS 洪寶惜 111005520
THE ART NEWSPAPER �AN AI BOT HAS FIGURED OUT HOW TO DRAW LIKE BANKSY. AND IT'S UNCANNY
GANksy aims to produce images that bear resemblance to works by the UK's most famous street artist
To create these images, Round has used a type of computerised machine learning framework known as a GAN (generative adversarial network). This specific GAN was trained for five days using a portfolio of hundreds of images of (potentially) Banksy's work, until it was able to produce an image that bears a superficial likeness to the originals.
作品論述 – GANKSY
Source : https://vole.wtf/ganksy/
作品論述 – BANKSY UKRAINE
reference : https://banksyexplained.com/a-great-british-spraycation-august-2021/
Banksy是一位匿名的英國塗鴉藝術家。他的街頭作品經常帶有諷刺意味，在旁則附有一些顛覆性、玩世不恭的黑色幽默和警世句子；其塗鴉大多運用獨特的模板技術拓印而成。他的作品富有濃厚政治風格，儼如一種以藝術方式表達的社會評論，並已經在世界各地不同城市的街道、牆壁與橋梁出現，甚至成為當地引人入勝的城市面貌。 維基百科
BANKSY UKRAINE
Instagram帳號；34家藝術機構，136個藝術家，總粉絲量6622萬
BASIC IDEA OF GAN (GENERATIVE ADVERSARIAL NETWORK)
Powered by: http://mattya.github.io/chainer-DCGAN/
BASIC IDEA OF GAN (GENERATIVE ADVERSARIAL NETWORK)
Powered by: http://mattya.github.io/chainer-DCGAN/
BASIC IDEA OF GAN (迭代的次數越來越多,所生成的圖片會愈來愈清晰)
Source of training data: https://zhuanlan.zhihu.com/p/24767059
BASIC IDEA OF GAN ( 迭代的次數越來越多,所生成的圖片會愈來愈清晰 )
( GAN) 生成對抗網路模型及其應用
https://generated.photos/
( GAN) 生成對抗網路模型及其應用
( GAN) 生成對抗網路模型及其應用
• Image generation tasks
ALL KINDS OF GAN …
AI趨勢周報第216期：GOOGLE用拖拉式介面讓GAN更準確生成圖片
https://www.ithome.com.tw/news/156973?fbclid=IwAR3rEPYQetj7vSdyzn78eULigCidooTSmvvL-q0Uy_kLDViQ_DAGzi5_vaE
Google聯手馬克斯普朗克研究院、MIT研究員，共同開發一款可精準生成圖片的模型DragGAN，並打造一套互動式UI介面，來讓使用者在畫面中，點擊想要修改的任意點和目標點，來驅動模型生成新圖片。
團隊指出，如何控制GAN精確地產出圖片，一直是個難題，傳統方法是透過手動標註的訓練資料或3D模型，但這種作法缺乏彈性、精確性和通用性。於是，團隊利用一種較少人探討的方法，來強化對GAN生成品質的控制性，也就是用互動式介面，來拖拉圖片中的任意點，準確地達到目標點。
為實現這個目標，團隊開發出DragGAN，由2大部分組成，一是基於特徵的運動監督方法，來實現選定點至目標點位置的移動，另一是新式點追蹤方法，利用GAN鑑別器特性來維持移動位置的在地化。總而言之，透過DragGAN，使用者可精確控制像素位置，來變形圖片，來改變人類、動物、車輛、風景等圖的姿勢、形狀、表情和布局。團隊表示，經測試，DragGAN在圖像處理和點追蹤任務的表現都比現有方法更好，未來打算擴大納入3D物件的影像生成。
DRAG YOUR GAN: INTERACTIVE POINT-BASED MANIPULATION ON THE GENERATIVE IMAGE MANIFOLD
https://vcai.mpi-inf.mpg.de/projects/DragGAN/
ABSTRACT
https://vcai.mpi-inf.mpg.de/projects/DragGAN/
Synthesizing visual content that meets users' needs often requires flexible and precise controllability of the pose, shape, expression, and layout of the generated objects. Existing approaches gain controllability of generative adversarial networks (GANs) via manually annotated training data or a prior 3D model, which often lack flexibility, precision, and generality. In this work, we study a powerful yet much less explored way of controlling GANs, that is, to "drag" any points of the image to precisely reach target points in a user-interactive manner. To achieve this, we propose DragGAN, which consists of two main components including: 1) a feature-based motion supervision that drives the handle point to move towards the target position, and 2) a new point tracking approach that leverages the discriminative GAN features to keep localizing the position of the handle points. Through DragGAN, anyone can deform an image with precise control over where pixels go, thus manipulating the pose, shape, expression, and layout of diverse categories such as animals, cars, humans, landscapes, etc. As these manipulations are performed on the learned generative image manifold of a GAN, they tend to produce realistic outputs even for challenging scenarios such as hallucinating occluded content and deforming shapes that consistently follow the object's rigidity. Both qualitative and quantitative comparisons demonstrate the advantage of DragGAN over prior approaches in the tasks of image manipulation and point tracking. We also showcase the manipulation of real images through GAN inversion.
DEMO
感謝您的聆聽