Syntharch: Interactive Image Search with Attribute-Conditioned Synthesis

Abstract

the database as feedback options, Syntharch causes less confusion to the user. Further, I establish that the specific search method I propose performs similarly or better in comparison to the conventional approach.Overall, my thesis presents a new approach of interactive image search, proposes a specific implementation following that approach, and validates the hypotheses that guided the search approach as well as the implementation choices.

Cross Domain Visual Search with Feature Learning using Multi-stream Transformer-based Architectures

“…Closest to our suggested creative workflow is the work of Yu & Kovashka [56], where the generated images are used to gather user feedback and improve the retrieval result; their method however is only text-conditioned. Our method learns a cross-domain embedding for scene sketches and images that is improved upon by the addition of layout synthesis in the framework; furthermore we show that images generated from our layouts are comparable to synthesis-only models.…”