Autoencoding beyond pixels using a learned similarity metric

作者：Anders Larsen, Søren Kaae Sønderby, Hugo Larochelle, Ole Winther · 发表于：Technical University of Denmark, DTU Orbit (Technical University of Denmark, DTU) · 年份：2016 · 被引用次数：1146 · 研究领域：Generative Adversarial Networks and Image Synthesis、Advanced Image and Video Retrieval Techniques、Advanced Image Processing Techniques

We present an autoencoder that leverages learned representations to better measure similarities in data space. By combining a variational autoencoder (VAE) with a generative adversarial network (GAN) we can use learned feature representations in the GAN discriminator as basis for the VAE reconstruction objective. Thereby, we replace element-wise errors with feature-wise errors to better capture the data distribution while offering invariance towards e.g. translation. We apply our method to images of faces and show that it outperforms VAEs with element-wise similarity measures in terms of visual fidelity. Moreover, we show that the method learns an embedding in which high-level abstract visual features (e.g. wearing glasses) can be modified using simple arithmetic.