Recent successes in image synthesis are powered by large-scale diffusion models.
However, most methods are currently limited to either text-or image-conditioned generation
for synthesizing an entire image, texture transfer or inserting objects into a user-specified
region. In contrast, in this work we focus on synthesizing complex interactions (ie, an
articulated hand) with a given object. Given an RGB image of an object, we aim to
hallucinate plausible images of a human hand interacting with it. We propose a two step …

In the supplementary material, we provide more implementation details and more qualitative
results. We discuss the details of articulation-agnostic hand proxy and how to apply DDPM
loss in the image space for training the LayoutNet (Sec. A. 1). We also present ablations on
Content-Net (Sec. A. 2). We further show:(i) the paired data construction method being
robust, in Sec. A. 3,(ii) baseline implementations details in Sec. A. 4,(iii) details of integrating
our approach to scene-level affordance prediction in Sec. A. 5. Finally, we discuss the …