Are Images Indistinguishable to Humans Also Indistinguishable to Classifiers?

• URL: http://arxiv.org/abs/2405.18029v3
• Date: Fri, 11 Oct 2024 05:11:35 GMT
• Title: Are Images Indistinguishable to Humans Also Indistinguishable to Classifiers?
• Authors: Zebin You, Xinyu Zhang, Hanzhong Guo, Jingdong Wang, Chongxuan Li,
• Abstract summary: We show that, from the perspective of neural network-based classifiers, even advanced diffusion models are still far from this goal. Our methodology naturally serves as a diagnostic tool for diffusion models by analyzing specific features of generated data. It sheds light on the model autophagy disorder and offers insights into the use of generated data.
• Score: 39.31679737754048
• License:
• Abstract: The ultimate goal of generative models is to perfectly capture the data distribution. For image generation, common metrics of visual quality (e.g., FID) and the perceived truthfulness of generated images seem to suggest that we are nearing this goal. However, through distribution classification tasks, we reveal that, from the perspective of neural network-based classifiers, even advanced diffusion models are still far from this goal. Specifically, classifiers are able to consistently and effortlessly distinguish real images from generated ones across various settings. Moreover, we uncover an intriguing discrepancy: classifiers can easily differentiate between diffusion models with comparable performance (e.g., U-ViT-H vs. DiT-XL), but struggle to distinguish between models within the same family but of different scales (e.g., EDM2-XS vs. EDM2-XXL). Our methodology carries several important implications. First, it naturally serves as a diagnostic tool for diffusion models by analyzing specific features of generated data. Second, it sheds light on the model autophagy disorder and offers insights into the use of generated data: augmenting real data with generated data is more effective than replacing it.

Related papers

• DreamDA: Generative Data Augmentation with Diffusion Models [68.22440150419003]
  This paper proposes a new classification-oriented framework DreamDA. DreamDA generates diverse samples that adhere to the original data distribution by considering training images in the original data as seeds. In addition, since the labels of the generated data may not align with the labels of their corresponding seed images, we introduce a self-training paradigm for generating pseudo labels.
  arXiv  Detail & Related papers  (2024-03-19T15:04:35Z)
• Training Class-Imbalanced Diffusion Model Via Overlap Optimization [55.96820607533968]
  Diffusion models trained on real-world datasets often yield inferior fidelity for tail classes. Deep generative models, including diffusion models, are biased towards classes with abundant training images. We propose a method based on contrastive learning to minimize the overlap between distributions of synthetic images for different classes.
  arXiv  Detail & Related papers  (2024-02-16T16:47:21Z)
• Learned representation-guided diffusion models for large-image generation [58.192263311786824]
  We introduce a novel approach that trains diffusion models conditioned on embeddings from self-supervised learning (SSL) Our diffusion models successfully project these features back to high-quality histopathology and remote sensing images. Augmenting real data by generating variations of real images improves downstream accuracy for patch-level and larger, image-scale classification tasks.
  arXiv  Detail & Related papers  (2023-12-12T14:45:45Z)
• InfoDiffusion: Representation Learning Using Information Maximizing Diffusion Models [35.566528358691336]
  InfoDiffusion is an algorithm that augments diffusion models with low-dimensional latent variables. InfoDiffusion relies on a learning objective regularized with the mutual information between observed and hidden variables. We find that InfoDiffusion learns disentangled and human-interpretable latent representations that are competitive with state-of-the-art generative and contrastive methods.
  arXiv  Detail & Related papers  (2023-06-14T21:48:38Z)
• Intriguing properties of synthetic images: from generative adversarial networks to diffusion models [19.448196464632]
  It is important to gain insight into which image features better discriminate fake images from real ones. In this paper we report on our systematic study of a large number of image generators of different families, aimed at discovering the most forensically relevant characteristics of real and generated images.
  arXiv  Detail & Related papers  (2023-04-13T11:13:19Z)
• Your Diffusion Model is Secretly a Zero-Shot Classifier [90.40799216880342]
  We show that density estimates from large-scale text-to-image diffusion models can be leveraged to perform zero-shot classification. Our generative approach to classification attains strong results on a variety of benchmarks. Our results are a step toward using generative over discriminative models for downstream tasks.
  arXiv  Detail & Related papers  (2023-03-28T17:59:56Z)
• Traditional Classification Neural Networks are Good Generators: They are Competitive with DDPMs and GANs [104.72108627191041]
  We show that conventional neural network classifiers can generate high-quality images comparable to state-of-the-art generative models. We propose a mask-based reconstruction module to make semantic gradients-aware to synthesize plausible images. We show that our method is also applicable to text-to-image generation by regarding image-text foundation models.
  arXiv  Detail & Related papers  (2022-11-27T11:25:35Z)
• Fast Unsupervised Brain Anomaly Detection and Segmentation with Diffusion Models [1.6352599467675781]
  We propose a method based on diffusion models to detect and segment anomalies in brain imaging. Our diffusion models achieve competitive performance compared with autoregressive approaches across a series of experiments with 2D CT and MRI data.
  arXiv  Detail & Related papers  (2022-06-07T17:30:43Z)
• Characterizing and Avoiding Problematic Global Optima of Variational Autoencoders [28.36260646471421]
  Variational Auto-encoders (VAEs) are deep generative latent variable models. Recent work shows that traditional training methods tend to yield solutions that violate desiderata. We show that both issues stem from the fact that the global optima of the VAE training objective often correspond to undesirable solutions.
  arXiv  Detail & Related papers  (2020-03-17T15:14:25Z)

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.