mlsanityData to 7 Oct 2026

Inverting Visual Representations with Convolutional Networks

arXiv 1506.02753 ↗First posted 9 Jun 2015cs.NE cs.CV cs.LGcited by 168

Not on the map. Why?

Abstract (arXiv)

Feature representations, both hand-designed and learned ones, are often hard to analyze and interpret, even when they are extracted from visual data. We propose a new approach to study image representations by inverting them with an up-convolutional neural network. We apply the method to shallow representations (HOG, SIFT, LBP), as well as to deep networks. For shallow representations our approach provides significantly better reconstructions than existing methods, revealing that there is surprisingly rich information contained in these features. Inverting a deep network trained on ImageNet provides several insights into the properties of the feature representation learned by the network. Most strikingly, the colors and the rough contours of an image can be reconstructed from activations in higher network layers and even from the predicted class probabilities.

Key facts. Cited by 168 papers in the map (data through 7 Oct 2026). First posted on arXiv 9 Jun 2015.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 42; 2022: 32; 2023: 31; 2024: 26; 2025: 21; 2026: 16.

168 papers.

Sort:

By area:

  1. 2024Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Surveycited by 641
  2. 2021Few-shot Image Generation via Cross-domain Correspondencecited by 206
  3. 2021Projected GANs Converge Fastercited by 204
  4. 2023Text Embeddings Reveal (Almost) As Much As Textcited by 189
  5. 2023Squeeze, Recover and Relabel: Dataset Condensation at ImageNet Scale From A New Perspectivecited by 136
  6. 2021Salient ImageNet: How to discover spurious features in Deep Learning?cited by 114
  7. 2022What do Vision Transformers Learn? A Visual Explorationcited by 75
  8. 2023Language Model Inversioncited by 66
  9. 2022Rethinking Reconstruction Autoencoder-Based Out-of-Distribution Detectioncited by 59
  10. 2022Plug & Play Attacks: Towards Robust and Flexible Model Inversion Attackscited by 58
  11. 2021Exploiting Explanations for Model Inversion Attackscited by 48
  12. 2022Is Geometry Enough for Matching in Visual Localization?cited by 48
38 more
  1. 2022Orchestra: Unsupervised Federated Learning via Globally Consistent Clusteringcited by 45
  2. 2024Operationalizing a Threat Model for Red-Teaming Large Language Models (LLMs)cited by 44
  3. 2024Extracting Prompts by Inverting LLM Outputscited by 40
  4. 2021Biometrics: Trust, but Verifycited by 39
  5. 2024Logits of API-Protected LLMs Leak Proprietary Informationcited by 38
  6. 2021Defining and Quantifying the Emergence of Sparse Concepts in DNNscited by 36
  7. 2022ResSFL: A Resistance Transfer Framework for Defending Model Inversion Attack in Split Federated Learningcited by 36
  8. 2021Video Autoencoder: self-supervised disentanglement of static 3D structure and motioncited by 34
  9. 2022Explaining Deepfake Detection by Analysing Image Matchingcited by 34
  10. 2021A Little Robustness Goes a Long Way: Leveraging Robust Features for Targeted Transfer Attackscited by 33
  11. 2023Does a Neural Network Really Encode Symbolic Concepts?cited by 27
  12. 2021Refining activation downsampling with SoftPoolcited by 26
  13. 2021Visual Explanations from Spiking Neural Networks using Interspike Intervalscited by 26
  14. 2024Transferable Embedding Inversion Attack: Uncovering Privacy Risks in Text Embeddings without Model Queriescited by 25
  15. 2023Privacy-preserving Adversarial Facial Featurescited by 24
  16. 2024Privacy-Preserving Face Recognition Using Trainable Feature Subtractioncited by 24
  17. 2021NinjaDesc: Content-Concealing Visual Descriptors via Adversarial Learningcited by 21
  18. 2023A Survey of What to Share in Federated Learning: Perspectives on Model Utility, Privacy Leakage, and Communication Efficiencycited by 21
  19. 2022It's All In the Teacher: Zero-Shot Quantization Brought Closer to the Teachercited by 20
  20. 2023Strong but simple: A Baseline for Domain Generalized Dense Perception by CLIP-based Transfer Learningcited by 20
  21. 2021How Privacy-Preserving are Line Clouds? Recovering Scene Details from 3D Linescited by 19
  22. 2021Relation-aware Compositional Zero-shot Learning for Attribute-Object Pair Recognitioncited by 19
  23. 2022Quantifying the Knowledge in a DNN to Explain Knowledge Distillation for Classificationcited by 19
  24. 2021Interpretable deep-learning models to help achieve the Sustainable Development Goalscited by 18
  25. 2022Plug-In Inversion: Model-Agnostic Inversion for Vision with Data Augmentationscited by 18
  26. 2023Valid P-Value for Deep Learning-Driven Salient Regioncited by 17
  27. 2023DGC-GNN: Leveraging Geometry and Color Cues for Visual Descriptor-Free 2D-3D Matchingcited by 17
  28. 2021A Game-Theoretic Taxonomy of Visual Concepts in DNNscited by 16
  29. 2021Designing Counterfactual Generators using Deep Model Inversioncited by 15
  30. 2022Core Risk Minimization using Salient ImageNetcited by 15
  31. 2021Counterfactual Explanations for Oblique Decision Trees: Exact, Efficient Algorithmscited by 14
  32. 2022Revisiting Sparse Convolutional Model for Visual Recognitioncited by 13
  33. 2024What do we learn from inverting CLIP models?cited by 13
  34. 2021Interpreting Representation Quality of DNNs for 3D Point Cloud Processingcited by 12
  35. 2021Contrastive Feature Loss for Image Predictioncited by 12
  36. 2021Understanding Invariance via Feedforward Inversion of Discriminatively Trained Classifierscited by 11
  37. 2021Ex-Model: Continual Learning from a Stream of Trained Modelscited by 11
  38. 2022Measuring Representational Robustness of Neural Networks Through Shared Invariancescited by 11

50 most cited of 168 All 168 (JSON)