Given an image and a target modification (eg an image of the Eiffel tower and the text"
without people and at night-time"), Compositional Image Retrieval (CIR) aims to retrieve the …

Despite their widespread success, Text-to-Image models (T2I) still struggle to produce
images that are both aesthetically pleasing and faithful to the user's input text. We introduce …

Diffusion models that can generate high-quality data from randomly sampled Gaussian
noises have become the mainstream generative method in both academia and industry. Are …

Impressive advances in text-to-image (T2I) generative models have yielded a plethora of
high performing models which are able to generate aesthetically appealing, photorealistic …

Recent text-to-image (T2I) models have benefited from large-scale and high-quality data,
demonstrating impressive performance. However, these T2I models still struggle to produce …

In this paper, we introduce a model designed to improve the prediction of image-text
alignment, targeting the challenge of compositional understanding in current visual …

Text-to-Image (T2I) models have made significant advancements in recent years, but they
still struggle to accurately capture intricate details specified in complex compositional …

Latent diffusion models excel at producing high-quality images from text. Yet, concerns
appear about the lack of diversity in the generated imagery. To tackle this, we introduce …

Although text-to-image (T2I) models exhibit remarkable generation capabilities, they
frequently fail to accurately bind semantically related objects or attributes in the input …

Building world models that accurately and comprehensively represent the real world is the
utmost aspiration for conditional image generative models as it would enable their use as …