Editable Generative Adversarial Networks: Generating and Editing Faces Simultaneously

Abstract

We propose a novel framework for simultaneously generating and manipulating the face images with desired attributes. While the state-of-the-art attribute editing technique has achieved the impressive performance for creating realistic attribute effects, they only address the image editing problem, using the input image as the condition of model. Recently, several studies attempt to tackle both novel face generation and attribute editing problem using a single solution. However, their image quality is still unsatisfactory. Our goal is to develop a single unified model that can simultaneously create and edit high quality face images with desired attributes. A key idea of our work is that we decompose the image into the latent and attribute vector in low dimensional representation, and then utilize the GAN framework for mapping the low dimensional representation to the image. In this way, we can address both the generation and editing problem by learning the generator. For qualitative and quantitative evaluations, the proposed algorithm outperforms recent algorithms addressing the same problem. Also, we show that our model can achieve the competitive performance with the state-of-the-art attribute editing technique in terms of attribute editing quality.

Detecting Manipulations in Video

“…Of particular relevance has been the application of such approaches for the automatic synthesis of highly realistic videos with impressive results. Among them, a popular task with direct implications on the aims of InVID is face swapping, where networks are trained to replace human faces in videos with increasingly more convincing results [8,2]. Other tasks include image-to-image translation [3,26], where the model learns to convert images from one domain to another (e.g.…”

Learning to Conceal: A Deep Learning Based Method for Preserving Privacy and Avoiding Prejudice

“…Several previous works have developed different machine learning architectures with an attempted to model the age progress in images (Ramanathan and Chellappa 2006;Ramanathan and Chellappa 2008;Zhang, Song, and Qi 2017;Baek, Bang, and Shim 2018). Zhang et al (2017) developed a Conditional Adversarial AutoEncoder (CAAE) model that studies the facial features, and important parameters that appear in each age segment.…”

“…That is, the generator tries to deceive the discriminator into thinking that the the images created by the generator are original. Baek et al (2018) created a face editing tool that is based on GANs. However, GANs are not appropriate for our goal, since we do not intend to produce an image that looks are real as possible, but to preserve the original image wile concealing the properties we want to conceal.…”