Disentangled Unsupervised Image Translation via Restricted Information Flow
This addresses the challenge of disentangling shared and domain-specific attributes in image translation for computer vision applications, representing a novel method rather than an incremental improvement.
The paper tackles the problem of unsupervised image-to-image translation in the many-to-many setting by inferring domain-specific attributes from data without architectural biases, achieving consistently high manipulation accuracy across synthetic and natural datasets.
Unsupervised image-to-image translation methods aim to map images from one domain into plausible examples from another domain while preserving structures shared across two domains. In the many-to-many setting, an additional guidance example from the target domain is used to determine domain-specific attributes of the generated image. In the absence of attribute annotations, methods have to infer which factors are specific to each domain from data during training. Many state-of-art methods hard-code the desired shared-vs-specific split into their architecture, severely restricting the scope of the problem. In this paper, we propose a new method that does not rely on such inductive architectural biases, and infers which attributes are domain-specific from data by constraining information flow through the network using translation honesty losses and a penalty on the capacity of domain-specific embedding. We show that the proposed method achieves consistently high manipulation accuracy across two synthetic and one natural dataset spanning a wide variety of domain-specific and shared attributes.