How Fragile Are Training-Free AI-Generated Image Detectors? A Controlled Audit of Score Direction, Preprocessing, and Compression

• URL: http://arxiv.org/abs/2606.20488v1
• Date: Thu, 18 Jun 2026 17:05:51 GMT
• Title: How Fragile Are Training-Free AI-Generated Image Detectors? A Controlled Audit of Score Direction, Preprocessing, and Compression
• Abstract summary: We audit two representative training-free scores -- an autoencoder-reconstruction score (AEROBLADE-style) and a noise-perturbation feature-similarity score (RIGID-style)<n>The audit yields three cautionary findings.
• Score: 6.3999417080764225
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Training-free detectors of AI-generated images promise generator-agnostic deployment without classifier training, yet their reported numbers are rarely compared under a single controlled protocol. We audit two representative training-free scores -- an autoencoder-reconstruction score (AEROBLADE-style) and a noise-perturbation feature-similarity score (RIGID-style) -- plus a naive feature-kNN control, on a common 1,500-image GenImage-derived benchmark spanning seven generators and JPEG compression at quality 70 and 50. The audit yields three cautionary findings. (i) Implementation details masquerade as method differences: replacing the LPIPS backbone (AlexNet -> VGG-16) changes overall AUROC by +0.085, and switching between resize-to-512 and native-resolution preprocessing flips per-generator conclusions by up to 0.38 AUROC. (ii) Score direction is not a property of the method but of its hyperparameters: the RIGID-style score is inverted (AUROC < 0.5) on SD1.5 and Wukong at noise level sigma=0.05, recovers to >0.5 for every generator at sigma=0.01, and collapses to 0.15 at sigma=0.3. (iii) Dataset format bias inflates robustness claims: without unified re-encoding, AUROC under JPEG-50 exceeds the clean condition for the AlexNet-backbone reconstruction score; after bias correction the residual anomaly localizes to a single generator (BigGAN). The audited scores have complementary per-generator failure sets, but naive z-score fusion does not beat the best single score, indicating that exploiting complementarity requires direction-aware combination.

Related papers

• Architecture-Dependent Causal Transfer of Activation States Across Large Language Models [51.56484100374058]
  Internal language activation states can be transferred causally between different large language models.<n>Injecting projected activations into a target model during generation produces a statistically significant, preregistered causal effect.<n>We conclude that end-to-end activation-state transfer between LLMs, as currently implemented, is architecture-dependent rather than universal.
  arXiv  Detail & Related papers  (2026-08-17T09:53:27Z)
• Tokenizer Generator Coupling in Medical Image Generation [0.0]
  Latent medical image generators usually treat the tokenizer as fixed preprocessing.<n>We test whether this separation is valid in a controlled ChestMNIST study at 64x64.<n> rankings depend jointly on the tokenizer, generator, and sampler.
  arXiv  Detail & Related papers  (2026-08-07T18:57:09Z)
• Adversarial Deepfake Generation and an Investigation of Purification-Based Adversarial Detection [9.42962723864652]
  This paper describes the participation of team "Go To Germany" in the ImageCLEF 2026 Deepfake Detection and Generation Task.<n>For the image generation task, we employ FLUX.1-dev with PuLID for identity-preserving face synthesis.<n>For the image detection task, we combine two complementary detectors - SigLIP+DINOv2 for AI-generated images and GenD-DINOv3 for face manipulations.
  arXiv  Detail & Related papers  (2026-07-28T15:19:16Z)
• GenSyn10: A Multi-Generative AI Dataset For Benchmarking Image Classification [17.505968172455784]
  GenSyn10 is a CIFAR-10-aligned synthetic image dataset of 60,000 images (10 classes, 32$times$32, 50k/10k split) generated using three state-of-the-art models.<n>We evaluate 17 image classification models under a four-stage protocol: real-data baseline, zero-shot transfer, fine-tuning, and retention.<n>Despite a measurable domain gap, CIFAR-10-trained models achieve up to 96.86% zero-shot accuracy on GenSyn10, increasing to 99.88% after fine-tuning.
  arXiv  Detail & Related papers  (2026-07-10T09:47:16Z)
• How Much Capacity Does EEG Denoising Need? Ultra-Compact Networks reveal Benchmark Saturation and Metric-Utility Gap [0.0]
  Deep learning EEG denoising architectures have scaled from tens of thousands to tens of millions of parameters, yet no prior study has isolated model capacity as the experimental variable.<n>We address both gaps by fixing architecture, loss, data split, and training recipe while sweeping only channel width from 1.05K to 40.26K parameters in a minimal depthwiseseparable convolutional U-Net.
  arXiv  Detail & Related papers  (2026-06-07T12:17:25Z)
• Amplifying, Not Learning: Fine-Tuned AI Text Detectors Amplify a Pretrained Direction [51.56484100374058]
  Text detectors amplify a pretrained typicality axis; they do not construct an AI-vs-human boundary.<n>On raw encoders before any task supervision, projecting onto centroid(AI)-centroid(HC3) achieves NYT-vs-HC3 AUROC 0.806/0.944/0.834 across three architectures.<n>On RoBERTa-base, raw projection exceeds fine-tuning; on RoBERTa-base, full fine-tuning reduces discrimination below raw on both fluent-formal populations tested.
  arXiv  Detail & Related papers  (2026-05-20T19:08:38Z)
• A Comparative Study of Modern Object Detectors for Robust Apple Detection in Orchard Imagery [0.0]
  This study establishes a benchmark for single-class apple detection on the public AppleBBCH81 dataset.<n>On the validation split, YOLOv10n achieves the best strict localization performance.<n>YOLOv10n attains the highest F1-score, whereas RT-DETR-L achieves very high recall but low precision.
  arXiv  Detail & Related papers  (2026-04-11T02:56:42Z)
• VA-$π$: Variational Policy Alignment for Pixel-Aware Autoregressive Generation [65.22452273252428]
  VA-$$ is a post-training framework to optimize autoregressive visual generation.<n>It unifies pixel reconstruction and autoregressive modeling.<n>It reduces FID from 14.36 to 7.65 and improves IS from 86.55 to 116.70 on LlamaGen-XXL.
  arXiv  Detail & Related papers  (2025-12-22T18:54:30Z)
• CARE What Fails: Contrastive Anchored-REflection for Verifiable Multimodal [84.71254539482369]
  Group-relative reinforcement learning with verifiable rewards (RLVR) often wastes the most informative data it already has the failures.<n>We present CARE, a failure-centric post-training framework for multimodal reasoning that turns errors into supervision.<n> CARE improves accuracy and training smoothness while explicitly increasing the share of learning signal that comes from failures.
  arXiv  Detail & Related papers  (2025-12-22T16:34:21Z)
• I Detect What I Don't Know: Incremental Anomaly Learning with Stochastic Weight Averaging-Gaussian for Oracle-Free Medical Imaging [2.384534878752428]
  We introduce an unsupervised, oracle-free framework that incrementally expands a trusted set of normal samples without any anomaly labels.<n>A frozen pretrained vision backbone is augmented with tiny convolutional adapters, ensuring rapid domain adaptation with negligible computational overhead.<n>On COVID-CXR, ROC-AUC improves from 0.9489 to 0.9982; on Pneumonia CXR, ROC-AUC rises from 0.6834 to 0.8968; and on Brain MRI ND-5, ROC-AUC increases from 0.6041 to 0.7269.
  arXiv  Detail & Related papers  (2025-11-05T23:28:14Z)
• Guided Score identity Distillation for Data-Free One-Step Text-to-Image Generation [62.30570286073223]
  Diffusion-based text-to-image generation models have demonstrated the ability to produce images aligned with textual descriptions.<n>We introduce a data-free guided distillation method that enables the efficient distillation of pretrained Diffusion models without access to the real training data.<n>By exclusively training with synthetic images generated by its one-step generator, our data-free distillation method rapidly improves FID and CLIP scores, achieving state-of-the-art FID performance while maintaining a competitive CLIP score.
  arXiv  Detail & Related papers  (2024-06-03T17:44:11Z)
• SADet: Learning An Efficient and Accurate Pedestrian Detector [68.66857832440897]
  This paper proposes a series of systematic optimization strategies for the detection pipeline of one-stage detector. It forms a single shot anchor-based detector (SADet) for efficient and accurate pedestrian detection. Though structurally simple, it presents state-of-the-art result and real-time speed of $20$ FPS for VGA-resolution images.
  arXiv  Detail & Related papers  (2020-07-26T12:32:38Z)
• ADAHESSIAN: An Adaptive Second Order Optimizer for Machine Learning [91.13797346047984]
  We introduce ADAHESSIAN, a second order optimization algorithm which dynamically incorporates the curvature of the loss function via ADAptive estimates. We show that ADAHESSIAN achieves new state-of-the-art results by a large margin as compared to other adaptive optimization methods.
  arXiv  Detail & Related papers  (2020-06-01T05:00:51Z)

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.