Neural Photo Editing with Introspective Adversarial Networks
This work addresses the problem of realistic image editing for applications in computer vision and graphics, representing an incremental advancement by hybridizing existing methods.
The paper tackled the challenge of making large, semantically coherent edits to existing images using generative models, and introduced the Neural Photo Editor with an Introspective Adversarial Network that achieved high visual fidelity in samples and reconstructions on datasets like CelebA, SVHN, and CIFAR-100.
The increasingly photorealistic sample quality of generative image models suggests their feasibility in applications beyond image generation. We present the Neural Photo Editor, an interface that leverages the power of generative neural networks to make large, semantically coherent changes to existing images. To tackle the challenge of achieving accurate reconstructions without loss of feature quality, we introduce the Introspective Adversarial Network, a novel hybridization of the VAE and GAN. Our model efficiently captures long-range dependencies through use of a computational block based on weight-shared dilated convolutions, and improves generalization performance with Orthogonal Regularization, a novel weight regularization method. We validate our contributions on CelebA, SVHN, and CIFAR-100, and produce samples and reconstructions with high visual fidelity.