Attribute-Driven Spontaneous Motion in Unpaired Image Translation
This work addresses the problem of incorporating geometric transformations in image translation for researchers and practitioners in computer vision, though it appears incremental as it builds on existing unpaired translation frameworks.
The paper tackles the limitation of current image translation methods in handling geometric transformations by proposing a spontaneous motion estimation module with refinement to learn attribute-driven deformation between domains. The approach achieves promising results in unpaired image translation tasks and enables applications based on spontaneous motion.
Current image translation methods, albeit effective to produce high-quality results in various applications, still do not consider much geometric transform. We in this paper propose the spontaneous motion estimation module, along with a refinement part, to learn attribute-driven deformation between source and target domains. Extensive experiments and visualization demonstrate effectiveness of these modules. We achieve promising results in unpaired-image translation tasks, and enable interesting applications based on spontaneous motion.