mlsanityData to 14 Oct 2026

Learned Smartphone ISP on Mobile GPUs with Deep Learning, Mobile AI & AIM 2022 Challenge: Report

arXiv 2211.03885 ↗First posted 7 Nov 2022cs.CV eess.IVcited by 13

Not on the map. Why?

Abstract (arXiv)

The role of mobile cameras increased dramatically over the past few years, leading to more and more research in automatic image quality enhancement and RAW photo processing. In this Mobile AI challenge, the target was to develop an efficient end-to-end AI-based image signal processing (ISP) pipeline replacing the standard mobile ISPs that can run on modern smartphone GPUs using TensorFlow Lite. The participants were provided with a large-scale Fujifilm UltraISP dataset consisting of thousands of paired photos captured with a normal mobile camera sensor and a professional 102MP medium-format FujiFilm GFX100 camera. The runtime of the resulting models was evaluated on the Snapdragon's 8 Gen 1 GPU that provides excellent acceleration results for the majority of common deep learning ops. The proposed solutions are compatible with all recent mobile GPUs, being able to process Full HD photos in less than 20-50 milliseconds while achieving high fidelity results. A detailed description of all models developed in this challenge is provided in this paper.

Key facts. Cited by 13 papers in the map (data through 14 Oct 2026). Builds on 55 of its 80 references that are in the map. First posted on arXiv 7 Nov 2022.

Builds on

55 of 80 references.

References · 55

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  2. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
  3. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,403
  4. 2015Domain-Adversarial Training of Neural Networkscited by 3,044
  5. 2018CBAM: Convolutional Block Attention Modulecited by 2,477
  6. 2019Searching for MobileNetV3cited by 1,994
  7. 2014Unsupervised Domain Adaptation by Backpropagationcited by 1,917
  8. 2017Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inferencecited by 1,593
  9. 2017Enhanced Deep Residual Networks for Single Image Super-Resolutioncited by 1,552
  10. 2016Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learningcited by 1,512
  11. 2018Image Super-Resolution Using Very Deep Residual Channel Attention Networkscited by 1,264
  12. 2018MnasNet: Platform-Aware Neural Architecture Search for Mobilecited by 943
38 more
  1. 2018Regularized Evolution for Image Classifier Architecture Searchcited by 909
  2. 2021RepVGG: Making VGG-style ConvNets Great Againcited by 522
  3. 2018FBNet: Hardware-Aware Efficient ConvNet Design via Differentiable Neural Architecture Searchcited by 485
  4. 2020ZeroQ: A Novel Zero Shot Quantization Frameworkcited by 269
  5. 2018Bi-Real Net: Enhancing the Performance of 1-bit CNNs With Improved Representational Capability and Advanced Training Algorithmcited by 235
  6. 2019MetaPruning: Meta Learning for Automatic Neural Network Channel Pruningcited by 227
  7. 2017DSLR-Quality Photos on Mobile Devices with Deep Convolutional Networkscited by 156
  8. 2019Quantization Networkscited by 153
  9. 2020FBNetV2: Differentiable Neural Architecture Search for Spatial and Channel Dimensionscited by 152
  10. 2020Residual Feature Distillation Network for Lightweight Image Super-Resolutioncited by 122
  11. 2021Learning Accurate Dense Correspondences and When to Trust Themcited by 113
  12. 2020Replacing Mobile Camera ISP with a Single Deep Learning Modelcited by 108
  13. 2019Trained Quantization Thresholds for Accurate and Efficient Fixed-Point Inference of Deep Neural Networkscited by 90
  14. 2020ResRep: Lossless CNN Pruning via Decoupling Remembering and Forgettingcited by 85
  15. 2020NTIRE 2020 Challenge on Real-World Image Super-Resolution: Methods and Resultscited by 83
  16. 2019AI Benchmark: All About Deep Learning on Smartphones in 2019cited by 77
  17. 2018AI Benchmark: Running Deep Neural Networks on Android Smartphonescited by 75
  18. 2019Unsupervised Learning for Real-World Super-Resolutioncited by 64
  19. 2020HistoGAN: Controlling Colors of GAN-Generated and Real Images via Color Histogramscited by 64
  20. 2019Learning Filter Basis for Convolutional Neural Network Compressioncited by 57
  21. 2020Rendering Natural Camera Bokeh Effect with Deep Learningcited by 57
  22. 2020AIM 2020 Challenge on Learned Image Signal Processing Pipelinecited by 44
  23. 2022Model-Based Image Signal Processors via Learnable Dictionariescited by 43
  24. 2021RepMLP: Re-parameterizing Convolutions into Fully-connected Layers for Image Recognitioncited by 40
  25. 2018PIRM Challenge on Perceptual Image Enhancement on Smartphones: Reportcited by 38
  26. 2020AWNet: Attentive Wavelet Network for Image ISPcited by 38
  27. 2021Learned Smartphone ISP on Mobile NPUs with Deep Learning, Mobile AI 2021 Challenge: Reportcited by 34
  28. 2021Real-Time Quantized Image Super-Resolution on Mobile NPUs, Mobile AI 2021 Challenge: Reportcited by 34
  29. 2017WESPE: Weakly Supervised Photo Enhancer for Digital Camerascited by 32
  30. 2019On-Device Neural Net Inference with Mobile GPUscited by 32
  31. 2020AIM 2020 Challenge on Rendering Realistic Bokehcited by 26
  32. 2020T-Basis: a Compact Representation for Neural Networkscited by 21
  33. 2022AIM 2022 Challenge on Super-Resolution of Compressed Image and Video: Dataset, Methods and Resultscited by 20
  34. 2021Fast and Accurate Single-Image Depth Estimation on Mobile Devices, Mobile AI 2021 Challenge: Reportcited by 19
  35. 2021Real-Time Video Super-Resolution on Smartphones with Deep Learning, Mobile AI 2021 Challenge: Reportcited by 19
  36. 2020Deploying Image Deblurring across Mobile Devices: A Perspective of Quality and Latencycited by 16
  37. 2020Controlling Information Capacity of Binary Neural Networkcited by 15
  38. 2021Fast Camera Image Denoising on Mobile GPUs with Deep Learning, Mobile AI 2021 Challenge: Reportcited by 15

50 most cited of 55 All 55 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 1; 2024: 4; 2025: 5; 2026: 3.

13 papers.

Sort:

  1. 2025NTIRE 2025 Challenge on Efficient Burst HDR and Restoration: Datasets, Methods, and Resultscited by 14
  2. 2025MobileIE: An Extremely Lightweight and Effective ConvNet for Real-Time Image Enhancement on Mobile Devicescited by 12
  3. 2023SYENet: A Simple Yet Effective Network for Multiple Low-Level Vision Tasks with Real-time Performance on Mobile Devicecited by 5
  4. 2025RAW Image Reconstruction from RGB on Smartphones. NTIRE 2025 Challenge Reportcited by 4
  5. 2025Learned Lightweight Smartphone ISP with Unpaired Datacited by 3
  6. 2024PQDynamicISP: Dynamically Controlled Image Signal Processor for Any Image Sensors Pursuing Perceptual Qualitycited by 2
  7. 2024Simple Image Signal Processing using Global Context Guidancecited by 2
  8. 2026Addressing Image Authenticity When Cameras Use Generative AIcited by 1
  9. 2024Realizing Unaligned Block-wise Pruning for DNN Acceleration on Mobile Devicescited by 0
  10. 2024Pushing Joint Image Denoising and Classification to the Edgecited by 0
  11. 2025Spatial-Frequency Aware for Object Detection in RAW Imagecited by 0
  12. 2026LatentBurst: A Fast and Efficient Multi Frame Super-Resolution for Hexadeca-Bayer Pattern CIS imagescited by 0
1 more
  1. 2026Benchmarking RAW and RGB Restoration in Image Signal Processorscited by 0