1 paper
Julien Boudier, Hugo Caselles-Dupré
While modern text-to-image diffusion models generate high-fidelity images, they offer limited control over the spatial and geometric structure of the output. To address this, we in…