mlsanityData to 14 Oct 2026

MobileIE: An Extremely Lightweight and Effective ConvNet for Real-Time Image Enhancement on Mobile Devices

arXiv 2507.01838 ↗First posted 2 Jul 2025cs.CVcited by 12

Not on the map. Why?

Abstract (arXiv)

Recent advancements in deep neural networks have driven significant progress in image enhancement (IE). However, deploying deep learning models on resource-constrained platforms, such as mobile devices, remains challenging due to high computation and memory demands. To address these challenges and facilitate real-time IE on mobile, we introduce an extremely lightweight Convolutional Neural Network (CNN) framework with around 4K parameters. Our approach integrates reparameterization with an Incremental Weight Optimization strategy to ensure efficiency. Additionally, we enhance performance with a Feature Self-Transform module and a Hierarchical Dual-Path Attention mechanism, optimized with a Local Variance-Weighted loss. With this efficient framework, we are the first to achieve real-time IE inference at up to 1,100 frames per second (FPS) while delivering competitive image quality, achieving the best trade-off between speed and performance across multiple IE tasks. The code will be available at https://github.com/AVC2-UESTC/MobileIE.git.

Key facts. Cited by 12 papers in the map (data through 14 Oct 2026). Builds on 77 of its 105 references that are in the map. First posted on arXiv 2 Jul 2025.

Builds on

77 of 105 references.

References · 77

  1. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,403
  2. 2017Squeeze-and-Excitation Networkscited by 4,789
  3. 2018CBAM: Convolutional Block Attention Modulecited by 2,477
  4. 2017ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devicescited by 1,200
  5. 2020Pre-Trained Image Processing Transformercited by 893
  6. 2018Deep Retinex Decomposition for Low-Light Enhancementcited by 723
  7. 2022Simple Baselines for Image Restorationcited by 708
  8. 2019ECA-Net: Efficient Channel Attention for Deep Convolutional Neural Networkscited by 575
  9. 2020Zero-Reference Deep Curve Estimation for Low-Light Image Enhancementcited by 542
  10. 2021RepVGG: Making VGG-style ConvNets Great Againcited by 522
  11. 2019GhostNet: More Features from Cheap Operationscited by 410
  12. 2018Efficient Attention: Attention with Linear Complexitiescited by 368
38 more
  1. 2021ResMLP: Feedforward networks for image classification with data-efficient trainingcited by 339
  2. 2020Learning Enriched Features for Real Image Restoration and Enhancementcited by 312
  3. 2020Retinex-inspired Unrolling with Cooperative Prior Architecture Search for Low-light Image Enhancementcited by 303
  4. 2023DiffIR: Efficient Diffusion Model for Image Restorationcited by 255
  5. 2019An Underwater Image Enhancement Benchmark Dataset and Beyondcited by 253
  6. 2017A General and Adaptive Robust Loss Functioncited by 236
  7. 2023Retinexformer: One-stage Retinex-based Transformer for Low-light Image Enhancementcited by 231
  8. 2021Learning to Enhance Low-Light Image via Zero-Reference Deep Curve Estimationcited by 192
  9. 2019Fast Underwater Image Enhancement for Improved Visual Perceptioncited by 183
  10. 2019ACNet: Strengthening the Kernel Skeletons for Powerful CNN via Asymmetric Convolution Blockscited by 178
  11. 2022Ultra-High-Definition Low-Light Image Enhancement: A Benchmark and Transformer-Based Methodcited by 174
  12. 2020Learning Image-adaptive 3D Lookup Tables for High Performance Photo Enhancement in Real-timecited by 163
  13. 2023Run, Don't Walk: Chasing Higher FLOPS for Faster Neural Networkscited by 130
  14. 2023Low-Light Image Enhancement with Wavelet-based Diffusion Modelscited by 127
  15. 2023RepViT: Revisiting Mobile CNN From ViT Perspectivecited by 123
  16. 2023Diff-Retinex: Rethinking Low-light Image Enhancement with A Generative Diffusion Modelcited by 118
  17. 2022MobileOne: An Improved One millisecond Mobile Backbonecited by 109
  18. 2020Replacing Mobile Camera ISP with a Single Deep Learning Modelcited by 108
  19. 2023DEA-Net: Single image dehazing based on detail-enhanced convolution and content-guided attentioncited by 91
  20. 2023Wavelet-based Fourier Information Interaction with Frequency Diffusion Adjustment for Underwater Image Restorationcited by 88
  21. 2021Diverse Branch Block: Building a Convolution as an Inception-like Unitcited by 87
  22. 2022You Only Need 90K Parameters to Adapt Light: A Light Weight Transformer for Image Enhancement and Exposure Correctioncited by 81
  23. 2018AI Benchmark: Running Deep Neural Networks on Android Smartphonescited by 75
  24. 2021Semantic-Guided Zero-Shot Learning for Low-Light Image/Video Enhancementcited by 60
  25. 2024Retinexmamba: Retinex-based Mamba for Low-light Image Enhancementcited by 59
  26. 2024The Ninth NTIRE 2024 Efficient Super-Resolution Challenge Reportcited by 57
  27. 2024Rewrite the Starscited by 55
  28. 2022Uncertainty Inspired Underwater Image Enhancementcited by 54
  29. 2020AIM 2020 Challenge on Learned Image Signal Processing Pipelinecited by 44
  30. 2022GhostNetV2: Enhance Cheap Operation with Long-Range Attentioncited by 44
  31. 2021Learning RAW-to-sRGB Mappings with Inaccurately Aligned Supervisioncited by 43
  32. 2023Efficient Multi-Scale Attention Module with Cross-Spatial Learningcited by 41
  33. 2024Correlation Matching Transformation Transformers for UHD Image Restorationcited by 39
  34. 2020AWNet: Attentive Wavelet Network for Image ISPcited by 38
  35. 2023Rethinking Mobile Block for Efficient Attention-based Modelscited by 36
  36. 2021Shallow-UWnet : Compressed Model for Underwater Image Enhancementcited by 35
  37. 2021Learned Smartphone ISP on Mobile NPUs with Deep Learning, Mobile AI 2021 Challenge: Reportcited by 34
  38. 2023Underwater Image Enhancement by Transformer-based Diffusion Model with Non-uniform Sampling for Skip Strategycited by 33

50 most cited of 77 All 77 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 4; 2026: 8.

12 papers.

  1. 2025StereoAdapter: Adapting Stereo Depth Estimation to Underwater Scenescited by 9
  2. 2025Towards Lightest Low-Light Image Enhancement Architecture for Mobile Devicescited by 6
  3. 2025IGDNet: Zero-Shot Robust Underexposed Image Enhancement via Illumination-Guided and Denoisingcited by 4
  4. 2026Revisiting Lightweight Low-Light Image Enhancement: From a YUV Color Space Perspectivecited by 1
  5. 2026DRNet: All-in-One Image Restoration via Prior-Guided Dynamic Reparameterizationcited by 1
  6. 2025JND-Guided Light-Weight Neural Pre-Filter for Perceptual Image Codingcited by 0
  7. 2026Advancing Visual Reliability: Color-Accurate Underwater Image Enhancement for Real-Time Underwater Missionscited by 0
  8. 2026FastSHADE: Fast Self-augmented Hierarchical Asymmetric Denoising for Efficient inference on mobile devicescited by 0
  9. 2026Bridging the Training-Deployment Gap: Gated Encoding and Multi-Scale Refinement for Efficient Quantization-Aware Image Enhancementcited by 0
  10. 2026NTIRE 2026 Challenge on Efficient Low Light Image Enhancement: Methods and Resultscited by 0
  11. 2026AnyMod-LLVE: Low-Light Video Enhancement with Modality-Agnostic Inferencecited by 0
  12. 2026Real-Time Underwater Image Enhancement via Frequency-Guided Dual-Path Attentioncited by 0