mlsanityData to 7 Oct 2026

Adaptive Convolutions with Per-pixel Dynamic Filter Atom

arXiv 2108.07895 ↗First posted 17 Aug 2021cs.CVcited by 11

Not on the map. Why?

Abstract (arXiv)

Applying feature dependent network weights have been proved to be effective in many fields. However, in practice, restricted by the enormous size of model parameters and memory footprints, scalable and versatile dynamic convolutions with per-pixel adapted filters are yet to be fully explored. In this paper, we address this challenge by decomposing filters, adapted to each spatial position, over dynamic filter atoms generated by a light-weight network from local features. Adaptive receptive fields can be supported by further representing each filter atom over sets of pre-fixed multi-scale bases. As plug-and-play replacements to convolutional layers, the introduced adaptive convolutions with per-pixel dynamic atoms enable explicit modeling of intra-image variance, while avoiding heavy computation, parameters, and memory cost. Our method preserves the appealing properties of conventional convolutions as being translation-equivariant and parametrically efficient. We present experiments to show that, the proposed method delivers comparable or even better performance across tasks, and are particularly effective on handling tasks with significant intra-image variance.

Key facts. Cited by 11 papers in the map (data through 7 Oct 2026). Builds on 37 of its 48 references that are in the map. First posted on arXiv 17 Aug 2021.

Builds on

37 of 48 references.

References · 37

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2019Searching for MobileNetV3cited by 1,984
  3. 2016Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Networkcited by 1,898
  4. 2017Enhanced Deep Residual Networks for Single Image Super-Resolutioncited by 1,543
  5. 2017Deformable Convolutional Networkscited by 1,481
  6. 2016Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoisingcited by 1,352
  7. 2018Image Super-Resolution Using Very Deep Residual Channel Attention Networkscited by 1,256
  8. 2013Spectral Networks and Locally Connected Networks on Graphscited by 1,097
  9. 2016HyperNetworkscited by 1,003
  10. 2015Accurate Image Super-Resolution Using Very Deep Convolutional Networkscited by 902
  11. 2018Deformable ConvNets v2: More Deformable, Better Resultscited by 666
  12. 2019Toward Real-World Single Image Super-Resolution: A New Benchmark and A New Modelcited by 400
25 more
  1. 2019Dynamic Convolution: Attention over Convolution Kernelscited by 355
  2. 2016Dynamic Filter Networkscited by 318
  3. 2018Toward Convolutional Blind Denoising of Real Photographscited by 274
  4. 2019CondConv: Conditionally Parameterized Convolutions for Efficient Inferencecited by 273
  5. 2018Composition Loss for Counting, Density Map Estimation and Localization in Dense Crowdscited by 186
  6. 2019Bayesian Loss for Crowd Count Estimation with Point Supervisioncited by 162
  7. 2016Using Fast Weights to Attend to the Recent Pastcited by 161
  8. 2017Burst Denoising with Kernel Prediction Networkscited by 156
  9. 2018Context-Aware Crowd Countingcited by 150
  10. 2019Learning from Synthetic Data for Crowd Counting in the Wildcited by 148
  11. 2019Pixel-Adaptive Convolutional Neural Networkscited by 126
  12. 2017Few-Shot Image Recognition by Predicting Parameters from Activationscited by 122
  13. 2019Variational Denoising Network: Toward Blind Noise Modeling and Removalcited by 117
  14. 2018Meta-Learning Probabilistic Inference For Predictioncited by 108
  15. 2017Switching Convolutional Neural Network for Crowd Countingcited by 93
  16. 2018Differentiable plasticity: training plastic neural networks with backpropagationcited by 83
  17. 2020Unified Dynamic Convolutional Network for Super-Resolution with Variational Degradationscited by 61
  18. 2020Learning Associative Inference Using Fast Weight Memorycited by 51
  19. 2018Revisiting Perspective Information for Efficient Crowd Countingcited by 49
  20. 2018Iterative Crowd Countingcited by 48
  21. 2020Backpropamine: training self-modifying neural networks with differentiable neuromodulated plasticitycited by 39
  22. 2018DCFNet: Deep Neural Network with Decomposed Convolutional Filterscited by 23
  23. 2019Stochastic Conditional Generative Networks with Basis Decompositioncited by 9
  24. 2017Super-Resolution with Deep Adaptive Image Resamplingcited by 3
  25. 2018In Defense of Single-column Networks for Crowd Countingcited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 3; 2023: 2; 2024: 3; 2025: 2; 2026: 1.

11 papers.

  1. 2023HyperE2VID: Improving Event-Based Video Reconstruction via Hypernetworkscited by 37
  2. 2024Tuning Timestep-Distilled Diffusion Model Using Pairwise Sample Optimizationcited by 25
  3. 2022APG: Adaptive Parameter Generation Network for Click-Through Rate Predictioncited by 12
  4. 2024Large Convolutional Model Tuning via Filter Subspacecited by 11
  5. 2023Decomposed Guided Dynamic Filters for Efficient RGB-Guided Depth Completioncited by 10
  6. 2022TVConv: Efficient Translation Variant Convolution for Layout-aware Visual Processingcited by 7
  7. 2022Fast and High-Quality Image Denoising via Malleable Convolutionscited by 6
  8. 2025Coeff-Tuning: A Graph Filter Subspace View for Tuning Attention-Based Large Modelscited by 2
  9. 2025Sparse Fine-Tuning of Transformers for Generative Taskscited by 1
  10. 2024Boosting Medical Image Segmentation Performance with Adaptive Convolution Layercited by 0
  11. 2026Structural Kolmogorov-Arnold Convolutions: Learnable Function on the Values or the Filter Shape as Parameter-Efficient Alternative to Per-Edge Convolutional KANscited by 0