Conditional Diffusion on Web-Scale Image Pairs leads to Diverse Image Variations

• URL: http://arxiv.org/abs/2405.14857v3
• Date: Wed, 02 Oct 2024 10:34:09 GMT
• Title: Conditional Diffusion on Web-Scale Image Pairs leads to Diverse Image Variations
• Authors: Manoj Kumar, Neil Houlsby, Emiel Hoogeboom,
• Abstract summary: Current image variation techniques involve adapting a text-to-image model to reconstruct an input image conditioned on the same image. We show that a diffusion model trained to reconstruct an input image from frozen embeddings, can reconstruct the image with minor variations. We propose a new pretraining strategy to generate image variations using a large collection of image pairs.
• Score: 32.892042877725125
• License:
• Abstract: Generating image variations, where a model produces variations of an input image while preserving the semantic context has gained increasing attention. Current image variation techniques involve adapting a text-to-image model to reconstruct an input image conditioned on the same image. We first demonstrate that a diffusion model trained to reconstruct an input image from frozen embeddings, can reconstruct the image with minor variations. Second, inspired by how text-to-image models learn from web-scale text-image pairs, we explore a new pretraining strategy to generate image variations using a large collection of image pairs. Our diffusion model \textit{Semantica} receives a random (encoded) image from a webpage as conditional input and denoises another noisy random image from the same webpage. We carefully examine various design choices for the image encoder, given its crucial role in extracting relevant context from the input image. Once trained, \textit{Semantica} can adaptively generate new images from a dataset by simply using images from that dataset as input. Finally, we identify limitations in standard image consistency metrics for evaluating image variations and propose alternative metrics based on few-shot generation.

Related papers

• Discriminative Class Tokens for Text-to-Image Diffusion Models [107.98436819341592]
  We propose a non-invasive fine-tuning technique that capitalizes on the expressive potential of free-form text. Our method is fast compared to prior fine-tuning methods and does not require a collection of in-class images. We evaluate our method extensively, showing that the generated images are: (i) more accurate and of higher quality than standard diffusion models, (ii) can be used to augment training data in a low-resource setting, and (iii) reveal information about the data used to train the guiding classifier.
  arXiv  Detail & Related papers  (2023-03-30T05:25:20Z)
• Masked and Adaptive Transformer for Exemplar Based Image Translation [16.93344592811513]
  Cross-domain semantic matching is challenging. We propose a masked and adaptive transformer (MAT) for learning accurate cross-domain correspondence. We devise a novel contrastive style learning method, for acquire quality-discriminative style representations.
  arXiv  Detail & Related papers  (2023-03-30T03:21:14Z)
• Unified Multi-Modal Latent Diffusion for Joint Subject and Text Conditional Image Generation [63.061871048769596]
  We present a novel Unified Multi-Modal Latent Diffusion (UMM-Diffusion) which takes joint texts and images containing specified subjects as input sequences. To be more specific, both input texts and images are encoded into one unified multi-modal latent space. Our method is able to generate high-quality images with complex semantics from both aspects of input texts and images.
  arXiv  Detail & Related papers  (2023-03-16T13:50:20Z)
• Uncovering the Disentanglement Capability in Text-to-Image Diffusion Models [60.63556257324894]
  A key desired property of image generative models is the ability to disentangle different attributes. We propose a simple, light-weight image editing algorithm where the mixing weights of the two text embeddings are optimized for style matching and content preservation. Experiments show that the proposed method can modify a wide range of attributes, with the performance outperforming diffusion-model-based image-editing algorithms.
  arXiv  Detail & Related papers  (2022-12-16T19:58:52Z)
• Cap2Aug: Caption guided Image to Image data Augmentation [41.53127698828463]
  Cap2Aug is an image-to-image diffusion model-based data augmentation strategy using image captions as text prompts. We generate captions from the limited training images and using these captions edit the training images using an image-to-image stable diffusion model. This strategy generates augmented versions of images similar to the training images yet provides semantic diversity across the samples.
  arXiv  Detail & Related papers  (2022-12-11T04:37:43Z)
• Vision Transformer Based Model for Describing a Set of Images as a Story [26.717033245063092]
  We propose a novel Vision Transformer Based Model for describing a set of images as a story. The proposed method extracts the distinct features of the input images using a Vision Transformer (ViT) The performance of our proposed model is evaluated using the Visual Story-Telling dataset (VIST)
  arXiv  Detail & Related papers  (2022-10-06T09:01:50Z)
• Retrieval Guided Unsupervised Multi-domain Image-to-Image Translation [59.73535607392732]
  Image to image translation aims to learn a mapping that transforms an image from one visual domain to another. We propose the use of an image retrieval system to assist the image-to-image translation task.
  arXiv  Detail & Related papers  (2020-08-11T20:11:53Z)
• Text as Neural Operator: Image Manipulation by Text Instruction [68.53181621741632]
  In this paper, we study a setting that allows users to edit an image with multiple objects using complex text instructions to add, remove, or change the objects. The inputs of the task are multimodal including (1) a reference image and (2) an instruction in natural language that describes desired modifications to the image. We show that the proposed model performs favorably against recent strong baselines on three public datasets.
  arXiv  Detail & Related papers  (2020-08-11T07:07:10Z)
• COCO-FUNIT: Few-Shot Unsupervised Image Translation with a Content Conditioned Style Encoder [70.23358875904891]
  Unsupervised image-to-image translation aims to learn a mapping of an image in a given domain to an analogous image in a different domain. We propose a new few-shot image translation model, COCO-FUNIT, which computes the style embedding of the example images conditioned on the input image. Our model shows effectiveness in addressing the content loss problem.
  arXiv  Detail & Related papers  (2020-07-15T02:01:14Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.