The Devil is in the Prompts: De-Identification Traces Enhance Memorization Risks in Synthetic Chest X-Ray Generation

• URL: http://arxiv.org/abs/2502.07516v2
• Date: Fri, 14 Feb 2025 17:24:56 GMT
• Title: The Devil is in the Prompts: De-Identification Traces Enhance Memorization Risks in Synthetic Chest X-Ray Generation
• Abstract summary: Generative models, particularly text-to-image (T2I) diffusion models, play a crucial role in medical image analysis.<n>These models are prone to training data memorization, posing significant risks to patient privacy.<n>This study presents the first systematic attempt to identify prompts that contribute the most to training data memorization.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Generative models, particularly text-to-image (T2I) diffusion models, play a crucial role in medical image analysis. However, these models are prone to training data memorization, posing significant risks to patient privacy. Synthetic chest X-ray generation is one of the most common applications in medical image analysis with the MIMIC-CXR dataset serving as the primary data repository for this task. This study presents the first systematic attempt to identify prompts and text tokens in MIMIC-CXR that contribute the most to training data memorization. Our analysis reveals two unexpected findings: (1) prompts containing traces of de-identification procedures (markers introduced to hide Protected Health Information) are the most memorized, and (2) among all tokens, de-identification markers contribute the most towards memorization. This highlights a broader issue with the standard anonymization practices and T2I synthesis with MIMIC-CXR. To exacerbate, existing inference-time memorization mitigation strategies are ineffective and fail to sufficiently reduce the model's reliance on memorized text tokens. On this front, we propose actionable strategies for different stakeholders to enhance privacy and improve the reliability of generative models in medical imaging. Finally, our results provide a foundation for future work on developing and benchmarking memorization mitigation techniques for synthetic chest X-ray generation using the MIMIC-CXR dataset. The anonymized code is available at https://anonymous.4open.science/r/diffusion_memorization-8011/

Related papers

• AnchorDiff: Topology-Aware Masked Diffusion with Confidence-based Rewriting for Radiology Report Generation [2.2748974006378933]
  Radiology report generation aims to automatically produce clinically accurate textual reports from medical images.<n>Existing methods rely on autoregressive (AR) language models, whose causal dependency structure restricts generation to a unidirectional left-to-right process.<n>We propose AnchorDiff, the first masked-diffusion framework for RRG that integrates knowledge-graph-derived clinical anchors into diffusion language modeling.
  arXiv  Detail & Related papers  (2026-05-16T16:42:43Z)
• Eliciting Medical Reasoning with Knowledge-enhanced Data Synthesis: A Semi-Supervised Reinforcement Learning Approach [62.0906177191353]
  Existing approaches distill chain-of-thought reasoning traces from large proprietary models via supervised fine-tuning, then conduct reinforcement learning (RL)<n>We propose MedSSR, a Medical Knowledge-enhanced data Synthesis and Semi-supervised Reinforcement learning framework.<n>Our framework first employs rare disease knowledge to synthesize distribution-controllable reasoning questions.<n>We then utilize the policy model itself to generate high-quality pseudo-labels. This enables a two-stage, intrinsic-to-extrinsic training paradigm: self-supervised RL on the pseudo-labeled data, followed by supervised RL on the human-annotated real data
  arXiv  Detail & Related papers  (2026-04-13T14:37:38Z)
• A Novel Metric for Detecting Memorization in Generative Models for Brain MRI Synthesis [4.16184304316315]
  DeepSSIM is a novel metric for quantifying memorization in generative models.<n>DeepSSIM achieves superior performance, improving F1 scores by an average of +52.03% over the best existing method.
  arXiv  Detail & Related papers  (2025-09-20T09:08:08Z)
• PriorRG: Prior-Guided Contrastive Pre-training and Coarse-to-Fine Decoding for Chest X-ray Report Generation [12.860257420677122]
  PriorRG is a chest X-ray report generation framework that emulates real-world clinical via a two-stage training pipeline.<n>In Stage 1, we introduce a prior-guided contrast pre-training scheme that leverages clinical context guidetemporal feature extraction.<n>In Stage 2, we integrate a prior-aware coarsetemporal decoding for generation that enhances prior knowledge with the vision encoders hidden states.
  arXiv  Detail & Related papers  (2025-08-07T13:02:20Z)
• CLARIFID: Improving Radiology Report Generation by Reinforcing Clinically Accurate Impressions and Enforcing Detailed Findings [1.515687944002438]
  We propose CLARIFID, a novel framework that directly optimize diagnostic correctness by mirroring the two-step workflow of experts.<n> CLARIFID learns the logical flow from Findings to Impression through section-aware pretraining.<n>We show that our method achieves superior clinical efficacy and outperforms existing baselines on both standard NLG metrics and clinically aware scores.
  arXiv  Detail & Related papers  (2025-07-23T05:57:59Z)
• Privacy-Preserving Chest X-ray Report Generation via Multimodal Federated Learning with ViT and GPT-2 [0.1874930567916036]
  The study proposes a Multimodal Federated Learning framework for chest X-ray report generation using the IU-Xray dataset.<n>The system utilizes a Vision Transformer (ViT) as the encoder and GPT-2 as the report generator, enabling decentralized training without sharing raw data.<n>The results show that FL can match or surpass centralized models in generating clinically relevant and semantically rich radiology reports.
  arXiv  Detail & Related papers  (2025-05-27T20:01:12Z)
• Enhanced Contrastive Learning with Multi-view Longitudinal Data for Chest X-ray Report Generation [15.257119888131609]
  We propose enhanced contrastive learning with Multi-view Longitudinal data to facilitate chest X-ray Report Generation, named MLRG. Specifically, we introduce a multi-view longitudinal contrast learning method that integrates spatial information from current multi-view images and temporal information from longitudinal data. We present a tokenized absence encoding technique to handle missing patient-specific prior knowledge, allowing the model to produce more accurate radiology reports based on available prior knowledge.
  arXiv  Detail & Related papers  (2025-02-27T12:59:04Z)
• ContextMRI: Enhancing Compressed Sensing MRI through Metadata Conditioning [51.26601171361753]
  We propose ContextMRI, a text-conditioned diffusion model for MRI that integrates granular metadata into the reconstruction process.<n>We show that increasing the fidelity of metadata, ranging from slice location and contrast to patient age, sex, and pathology, systematically boosts reconstruction performance.
  arXiv  Detail & Related papers  (2025-01-08T05:15:43Z)
• MRGen: Diffusion-based Controllable Data Engine for MRI Segmentation towards Unannotated Modalities [59.61465292965639]
  This paper investigates a new paradigm for leveraging generative models in medical applications.<n>We propose a diffusion-based data engine, termed MRGen, which enables generation conditioned on text prompts and masks.
  arXiv  Detail & Related papers  (2024-12-04T16:34:22Z)
• Detecting, Explaining, and Mitigating Memorization in Diffusion Models [49.438362005962375]
  We introduce a straightforward yet effective method for detecting memorized prompts by inspecting the magnitude of text-conditional predictions. Our proposed method seamlessly integrates without disrupting sampling algorithms, and delivers high accuracy even at the first generation step. Building on our detection strategy, we unveil an explainable approach that shows the contribution of individual words or tokens to memorization.
  arXiv  Detail & Related papers  (2024-07-31T16:13:29Z)
• Extracting Training Data from Unconditional Diffusion Models [76.85077961718875]
  diffusion probabilistic models (DPMs) are being employed as mainstream models for generative artificial intelligence (AI) We aim to establish a theoretical understanding of memorization in DPMs with 1) a memorization metric for theoretical analysis, 2) an analysis of conditional memorization with informative and random labels, and 3) two better evaluation metrics for measuring memorization. Based on the theoretical analysis, we propose a novel data extraction method called textbfSurrogate condItional Data Extraction (SIDE) that leverages a trained on generated data as a surrogate condition to extract training data directly from unconditional diffusion models.
  arXiv  Detail & Related papers  (2024-06-18T16:20:12Z)
• Memory-based Cross-modal Semantic Alignment Network for Radiology Report Generation [5.314122066634083]
  Key information related to disease accounts for a small proportion in both image and report. It is hard for the model to learn the latent relation between the radiology image and its report, thus failing to generate fluent and accurate radiology reports. We propose a memory-based cross-modal alignment model (MCSAM) following an encoder-decoder paradigm.
  arXiv  Detail & Related papers  (2024-03-31T07:30:41Z)
• MedPromptX: Grounded Multimodal Prompting for Chest X-ray Diagnosis [1.2903829793534272]
  Chest X-ray images are commonly used for predicting acute and chronic cardiopulmonary conditions.<n>Efforts to integrate them with structured clinical data face challenges due to incomplete electronic health records.<n>This paper introduces MedPromptX, the first clinical decision support system that integrates multimodal large language models (MLLMs), few-shot prompting (FP) and visual grounding (VG)<n>Results demonstrate the SOTA performance of MedPromptX, achieving an 11% improvement in F1-score compared to the baselines.
  arXiv  Detail & Related papers  (2024-03-22T19:19:51Z)
• Medical Image Captioning via Generative Pretrained Transformers [57.308920993032274]
  We combine two language models, the Show-Attend-Tell and the GPT-3, to generate comprehensive and descriptive radiology records. The proposed model is tested on two medical datasets, the Open-I, MIMIC-CXR, and the general-purpose MS-COCO.
  arXiv  Detail & Related papers  (2022-09-28T10:27:10Z)
• Longitudinal detection of new MS lesions using Deep Learning [0.0]
  We describe a deep-learning-based pipeline addressing the task of detecting and segmenting new MS lesions. First, we propose to use transfer-learning from a model trained on a segmentation task using single time-points. Second, we propose a data synthesis strategy to generate realistic longitudinal time-points with new lesions.
  arXiv  Detail & Related papers  (2022-06-16T16:09:04Z)
• Cross-modal Memory Networks for Radiology Report Generation [30.13916304931662]
  Cross-modal memory networks (CMN) are proposed to enhance the encoder-decoder framework for radiology report generation. Our model is able to better align information from radiology images and texts so as to help generating more accurate reports in terms of clinical indicators.
  arXiv  Detail & Related papers  (2022-04-28T02:32:53Z)
• MG-NET: Leveraging Pseudo-Imaging for Multi-Modal Metagenome Analysis [5.04905391284093]
  We propose MG-Net, a self-supervised representation learning framework. We show that MG-Net can learn robust representations from unlabeled data. Experiments show that the learned features outperform current baseline metagenome representations.
  arXiv  Detail & Related papers  (2021-07-21T05:53:01Z)
• METGAN: Generative Tumour Inpainting and Modality Synthesis in Light Sheet Microscopy [4.872960046536882]
  We introduce a novel generative method which leverages real anatomical information to generate realistic image-label pairs of tumours. We construct a dual-pathway generator, for the anatomical image and label, trained in a cycle-consistent setup, constrained by an independent, pretrained segmentor. The generated images yield significant quantitative improvement compared to existing methods.
  arXiv  Detail & Related papers  (2021-04-22T11:18:17Z)
• Many-to-One Distribution Learning and K-Nearest Neighbor Smoothing for Thoracic Disease Identification [83.6017225363714]
  deep learning has become the most powerful computer-aided diagnosis technology for improving disease identification performance. For chest X-ray imaging, annotating large-scale data requires professional domain knowledge and is time-consuming. In this paper, we propose many-to-one distribution learning (MODL) and K-nearest neighbor smoothing (KNNS) methods to improve a single model's disease identification performance.
  arXiv  Detail & Related papers  (2021-02-26T02:29:30Z)
• G-MIND: An End-to-End Multimodal Imaging-Genetics Framework for Biomarker Identification and Disease Classification [49.53651166356737]
  We propose a novel deep neural network architecture to integrate imaging and genetics data, as guided by diagnosis, that provides interpretable biomarkers. We have evaluated our model on a population study of schizophrenia that includes two functional MRI (fMRI) paradigms and Single Nucleotide Polymorphism (SNP) data.
  arXiv  Detail & Related papers  (2021-01-27T19:28:04Z)

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.