mlsanityData to 7 Oct 2026

Robust High-Resolution Video Matting with Temporal Guidance

arXiv 2108.11515 ↗First posted 25 Aug 2021cs.CVcited by 136

Not on the map. Why?

Abstract (arXiv)

We introduce a robust, real-time, high-resolution human video matting method that achieves new state-of-the-art performance. Our method is much lighter than previous approaches and can process 4K at 76 FPS and HD at 104 FPS on an Nvidia GTX 1080Ti GPU. Unlike most existing methods that perform video matting frame-by-frame as independent images, our method uses a recurrent architecture to exploit temporal information in videos and achieves significant improvements in temporal coherence and matting quality. Furthermore, we propose a novel training strategy that enforces our network on both matting and segmentation objectives. This significantly improves our model's robustness. Our method does not require any auxiliary inputs such as a trimap or a pre-captured background image, so it can be widely applied to existing human matting applications.

Key facts. Cited by 136 papers in the map (data through 7 Oct 2026). Builds on 24 of its 47 references that are in the map. First posted on arXiv 25 Aug 2021.

Builds on

24 of 47 references.

References · 24

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  3. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,441
  4. 2015Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shiftcited by 6,979
  5. 2018MobileNetV2: Inverted Residuals and Linear Bottleneckscited by 4,773
  6. 2014Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translationcited by 4,016
  7. 2018Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentationcited by 3,210
  8. 2017Rethinking Atrous Convolution for Semantic Image Segmentationcited by 2,154
  9. 2019Searching for MobileNetV3cited by 1,984
  10. 2015Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcastingcited by 1,428
  11. 2019PointRend: Image Segmentation as Renderingcited by 439
  12. 2020Real-Time High-Resolution Background Mattingcited by 172
12 more
  1. 2017Deep Image Mattingcited by 150
  2. 2015Delving Deeper into Convolutional Networks for Learning Video Representationscited by 146
  3. 2020Background Matting: The World is Your Green Screencited by 104
  4. 2019RVOS: End-to-End Recurrent Network for Video Object Segmentationcited by 88
  5. 2019Context-Aware Image Matting for Simultaneous Foreground and Alpha Estimationcited by 79
  6. 2017Learning Video Object Segmentation with Visual Memorycited by 66
  7. 2017Information-Flow Mattingcited by 59
  8. 2018Fast End-to-End Trainable Guided Filtercited by 50
  9. 2020F, B, Alpha Mattingcited by 50
  10. 2021Deep Video Matting via Spatio-Temporal Alignment and Aggregationcited by 41
  11. 2017Fast Deep Matting for Portrait Animation on Mobile Phonecited by 30
  12. 2019Semantic Segmentation of Video Sequences with Convolutional LSTMscited by 14

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 2; 2022: 24; 2023: 35; 2024: 26; 2025: 27; 2026: 22.

136 papers.

Sort:

By area:

  1. 2024SplattingAvatar: Realistic Real-Time Human Avatars with Mesh-Embedded Gaussian Splattingcited by 171
  2. 2023HumanRF: High-Fidelity Neural Radiance Fields for Humans in Motioncited by 161
  3. 2022Instant Volumetric Head Avatarscited by 157
  4. 2023Vid2Avatar: 3D Avatar Reconstruction from Videos in the Wild via Self-supervised Scene Decompositioncited by 134
  5. 2022NeRF-Art: Text-Driven Neural Radiance Fields Stylizationcited by 126
  6. 2022Highly Accurate Dichotomous Image Segmentationcited by 110
  7. 2023FlashAvatar: High-fidelity Head Avatar with Efficient Gaussian Embeddingcited by 107
  8. 2023EMDB: The Electromagnetic Database of Global 3D Human Pose and Shape in the Wildcited by 95
  9. 20234K4D: Real-Time 4D View Synthesis at 4K Resolutioncited by 91
  10. 2023Hi4D: 4D Instance Segmentation of Close Human Interactioncited by 87
  11. 2023MonoHuman: Animatable Human Neural Field from Monocular Videocited by 85
  12. 2022AvatarMAV: Fast 3D Head Avatar Reconstruction Using Motion-Aware Neural Voxelscited by 73
38 more
  1. 2022Harmonizer: Learning to Perform White-Box Image and Video Harmonizationcited by 61
  2. 2023HeadGaS: Real-Time Animatable Head Avatars via 3D Gaussian Splattingcited by 60
  3. 2022Neural Surface Reconstruction of Dynamic Scenes with Monocular RGB-D Cameracited by 57
  4. 2023ViTMatte: Boosting Image Matting with Pretrained Plain Vision Transformerscited by 56
  5. 2023MVHumanNet: A Large-scale Dataset of Multi-view Daily Dressing Human Capturescited by 54
  6. 2023StyleAvatar: Real-time Photo-realistic Portrait Avatar from a Single Videocited by 52
  7. 2024CAP4D: Creating Animatable 4D Portrait Avatars with Morphable Multi-View Diffusion Modelscited by 52
  8. 2023LatentAvatar: Learning Latent Expression Code for Expressive Neural Head Avatarcited by 50
  9. 2023AutoDecoding Latent 3D Diffusion Modelscited by 50
  10. 2024AnimateDiff-Lightning: Cross-Model Diffusion Distillationcited by 47
  11. 2023RenderMe-360: A Large Digital Asset Library and Benchmarks Towards High-fidelity Head Avatarscited by 45
  12. 2024GAF: Gaussian Avatar Reconstruction from Monocular Videos via Multi-view Diffusioncited by 36
  13. 2022SunStage: Portrait Reconstruction and Relighting using the Sun as a Light Stagecited by 30
  14. 2023Matting Anythingcited by 30
  15. 2022HARP: Personalized Hand Reconstruction from a Monocular RGB Videocited by 28
  16. 2023Relightable and Animatable Neural Avatar from Sparse-View Videocited by 28
  17. 2022FOF: Learning Fourier Occupancy Field for Monocular Real-time Human Reconstructioncited by 27
  18. 2025RGBAvatar: Reduced Gaussian Blendshapes for Online Modeling of Head Avatarscited by 27
  19. 2024Make-Your-Anchor: A Diffusion-based 2D Avatar Generation Frameworkcited by 26
  20. 2022Human Instance Matting via Mutual Guidance and Multi-Instance Refinementcited by 25
  21. 2025MatAnyone: Stable Video Matting with Consistent Memory Propagationcited by 24
  22. 2025HRAvatar: High-Quality and Relightable Gaussian Head Avatarcited by 24
  23. 2022Neural Capture of Animatable 3D Human from Monocular Videocited by 22
  24. 2022Global Spectral Filter Memory Network for Video Object Segmentationcited by 22
  25. 2022VMFormer: End-to-End Video Matting with Transformercited by 20
  26. 2024MaGGIe: Masked Guided Gradual Human Instance Mattingcited by 18
  27. 2024URHand: Universal Relightable Handscited by 17
  28. 2024PKU-DyMVHumans: A Multi-View Video Benchmark for High-Fidelity Dynamic Human Modelingcited by 17
  29. 2025MVP4D: Multi-View Portrait Video Diffusion for Animatable 4D Avatarscited by 17
  30. 2022Neural Parameterization for Dynamic Human Head Editingcited by 16
  31. 2023LumiGAN: Unconditional Generation of Relightable 3D Human Facescited by 14
  32. 2023GAN-Avatar: Controllable Personalized GAN-based Human Head Avatarcited by 14
  33. 2024Authentic Hand Avatar from a Phone Scan via Universal Hand Modelcited by 13
  34. 2025FVQ: A Large-Scale Dataset and an LMM-based Method for Face Video Quality Assessmentcited by 13
  35. 2022One-Trimap Video Mattingcited by 12
  36. 2025TransPixeler: Advancing Text-to-Video Generation with Transparencycited by 12
  37. 2026Video Generation Models are General-Purpose Vision Learnerscited by 12
  38. 2022Robust Human Matting via Semantic Guidancecited by 11

50 most cited of 136 All 136 (JSON)