Privacy Preserving Visual Question Answering
Not on the map. Why?
Abstract (arXiv)
We introduce a novel privacy-preserving methodology for performing Visual Question Answering on the edge. Our method constructs a symbolic representation of the visual scene, using a low-complexity computer vision model that jointly predicts classes, attributes and predicates. This symbolic representation is non-differentiable, which means it cannot be used to recover the original image, thereby keeping the original image private. Our proposed hybrid solution uses a vision model which is more than 25 times smaller than the current state-of-the-art (SOTA) vision models, and 100 times smaller than end-to-end SOTA VQA models. We report detailed error analysis and discuss the trade-offs of using a distilled vision model and a symbolic representation of the visual scene.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 23 of its 33 references that are in the map. First posted on arXiv 15 Feb 2022.
Builds on
23 of 33 references.
References · 23
- 2015Deep Residual Learning for Image Recognitioncited by 51,809
- 2014Microsoft COCO: Common Objects in Contextcited by 17,498
- 2015Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networkscited by 9,617
- 2019EfficientNet: Rethinking Model Scaling for Convolutional Neural Networkscited by 6,415
- 2016Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotationscited by 2,646
- 2015VQA: Visual Question Answeringcited by 2,489
- 2019ViLBERT: Pretraining Task-Agnostic Visiolinguistic Representations for Vision-and-Language Taskscited by 2,219
- 2020YOLOv4: Optimal Speed and Accuracy of Object Detectioncited by 1,624
- 2019LXMERT: Learning Cross-Modality Encoder Representations from Transformerscited by 1,548
- 2017Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answeringcited by 1,310
- 2020Oscar: Object-Semantics Aligned Pre-training for Vision-Language Taskscited by 1,213
- 2019EfficientDet: Scalable and Efficient Object Detectioncited by 1,018
11 more
- 2018The Open Images Dataset V4: Unified image classification, object detection, and visual relationship detection at scalecited by 542
- 2019Deep Modular Co-Attention Networks for Visual Question Answeringcited by 338
- 2017A Survey of Model Compression and Acceleration for Deep Neural Networkscited by 298
- 2020In Defense of Grid Features for Visual Question Answeringcited by 155
- 2017Learning to Reason: End-to-End Module Networks for Visual Question Answeringcited by 125
- 2016Visual Question Answering: A Survey of Methods and Datasetscited by 123
- 2018Pythia v0.1: the Winning Entry to the VQA Challenge 2018cited by 62
- 2019Language-Conditioned Graph Networks for Relational Reasoningcited by 55
- 2020MiniVLM: A Smaller and Faster Vision-Language Modelcited by 51
- 2020Deep Multimodal Neural Architecture Searchcited by 34
- 2019Bilinear Graph Networks for Visual Question Answeringcited by 22
Built on by
Citing papers per year: 2023: 1; 2024: 0; 2025: 0; 2026: 1.
2 papers.