Image-to-image Translation as a Unique Source of Knowledge
This work addresses a domain-specific issue for satellite imagery analysis, but it is incremental as it builds on existing I2I methods without introducing new paradigms.
The paper tackles the problem of unclear usability and transfer extent in image-to-image translation between dissimilar domains like SAR and optical satellite imagery, by translating labeled datasets and evaluating transfer, and finds that stacking models improves performance.
Image-to-image (I2I) translation is an established way of translating data from one domain to another but the usability of the translated images in the target domain when working with such dissimilar domains as the SAR/optical satellite imagery ones and how much of the origin domain is translated to the target domain is still not clear enough. This article address this by performing translations of labelled datasets from the optical domain to the SAR domain with different I2I algorithms from the state-of-the-art, learning from transferred features in the destination domain and evaluating later how much from the original dataset was transferred. Added to this, stacking is proposed as a way of combining the knowledge learned from the different I2I translations and evaluated against single models.