Beyond Blanket Masking: Examining Granularity for Privacy Protection in Images Captured by Blind and Low Vision Users
On the map: Multimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness
Abstract (arXiv)
As visual assistant systems powered by visual language models (VLMs) become more prevalent, concerns over user privacy have grown, particularly for blind and low vision users who may unknowingly capture personal private information in their images. Existing privacy protection methods rely on coarse-grained segmentation, which uniformly masks entire private objects, often at the cost of usability. In this work, we propose FiGPriv, a fine-grained privacy protection framework that selectively masks only high-risk private information while preserving low-risk information. Our approach integrates fine-grained segmentation with a data-driven risk scoring mechanism. We evaluate our framework using the BIV-Priv-Seg dataset and show that FiG-Priv preserves +26% of image content, enhancing the ability of VLMs to provide useful responses by 11% and identify the image content by 45%, while ensuring privacy protection. Project Page: https://artcs1.github.io/VLMPrivacy/
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Placed in MLLM jailbreaks and adversarial robustness, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 18 of its 45 references that are in the map. First posted on arXiv 12 Aug 2025.
Builds on
18 of 45 references.
References · 18
- 2023Segment Anythingcited by 8,234
- 2025Qwen2.5-VL Technical Reportcited by 5,587
- 2024GPT-4o System Cardcited by 4,588
- 2025Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilitiescited by 3,796
- 2022V2X-ViT: Vehicle-to-Everything Cooperative Perception with Vision Transformercited by 320
- 2022Where2comm: Communication-Efficient Collaborative Perception via Spatial Confidence Mapscited by 255
- 2018BOP: Benchmark for 6D Object Pose Estimationcited by 254
- 2022CoBEVT: Cooperative Bird's Eye View Semantic Segmentation with Sparse Transformerscited by 227
- 2017Towards a Visual Privacy Advisor: Understanding and Predicting Privacy Risks in Imagescited by 101
- 2018Privacy Preserving Machine Learning: Threats and Solutionscited by 79
- 2024EVF-SAM: Early Vision-Language Fusion for Text-Prompted Segment Anything Modelcited by 65
- 2021PGNet: Real-time Arbitrarily-Shaped Text Spotting with Point Gathering Networkcited by 49
6 more
- 2024Introducing HOT3D: An Egocentric Dataset for 3D Hand and Object Trackingcited by 48
- 2023CORE: Cooperative Reconstruction for Multi-Agent Perceptioncited by 41
- 2023QUEST: Query Stream for Practical Cooperative Perceptioncited by 23
- 2025CoopDETR: A Unified Cooperative Perception Framework for 3D Detection via Object Querycited by 9
- 2024BIV-Priv-Seg: Locating Private Content in Images Taken by People With Visual Impairmentscited by 5
- 2025Privacy Risk Predictions Based on Fundamental Understanding of Personal Data and an Evolving Threat Landscapecited by 1
Built on by
Citing papers per year: 2026: 2.
2 papers.
Similar, not cited
Close in wording, no citation either way
- 2024The Phantom Menace: Unmasking Privacy Leakages in Vision-Language Models
- 2026Rethinking Visual Privacy: A Compositional Privacy Risk Framework for Severity Assessment with VLMs
- 2025MultiPriv: Benchmarking Individual-Level Privacy Reasoning in Vision-Language Models
- 2025Who Can See Through You? Adversarial Shielding Against VLM-Based Attribute Inference Attacks
- 2025Assessing Visual Privacy Risks in Multimodal AI: A Novel Taxonomy-Grounded Evaluation of Vision-Language Models
- 2025VIP: Visual Information Protection through Adversarial Attacks on Vision-Language Models
- 2026Vision Language Model Helps Private Information De-Identification in Vision Data
- 2025On the Robustness of Vision-Language Models in Zero-shot Privacy Classification