Controllable Garment Transfer
This addresses a limitation in virtual try-on systems for users seeking customizable garment adjustments, though it appears incremental by adding tweaking functionality to existing models.
The paper tackles the problem of enabling on-the-fly garment modification in image-based garment transfer, allowing users to control attributes like sleeve length, waist width, and texture.
Image-based garment transfer replaces the garment on the target human with the desired garment; this enables users to virtually view themselves in the desired garment. To this end, many approaches have been proposed using the generative model and have shown promising results. However, most fail to provide the user with on the fly garment modification functionality. We aim to add this customizable option of "garment tweaking" to our model to control garment attributes, such as sleeve length, waist width, and garment texture.