The Spatially-Correlative Loss for Various Image Translation Tasks
This addresses the challenge of maintaining structural consistency in image translation for computer vision applications, representing an incremental advance by introducing a novel loss function.
The paper tackles the problem of preserving scene structure consistency during unpaired image-to-image translation across large domain gaps, proposing a spatially-correlative loss that shows distinct improvements over baseline models in single-modal, multi-modal, and single-image translation tasks.
We propose a novel spatially-correlative loss that is simple, efficient and yet effective for preserving scene structure consistency while supporting large appearance changes during unpaired image-to-image (I2I) translation. Previous methods attempt this by using pixel-level cycle-consistency or feature-level matching losses, but the domain-specific nature of these losses hinder translation across large domain gaps. To address this, we exploit the spatial patterns of self-similarity as a means of defining scene structure. Our spatially-correlative loss is geared towards only capturing spatial relationships within an image rather than domain appearance. We also introduce a new self-supervised learning method to explicitly learn spatially-correlative maps for each specific translation task. We show distinct improvement over baseline models in all three modes of unpaired I2I translation: single-modal, multi-modal, and even single-image translation. This new loss can easily be integrated into existing network architectures and thus allows wide applicability.