mlsanityData to 14 Oct 2026

DIFFUMA: High-Fidelity Spatio-Temporal Video Prediction via Dual-Path Mamba and Diffusion Enhancement

arXiv 2507.06738 ↗First posted 9 Jul 2025cs.CV cs.AIcited by 0

On the map: State-space sequence models (S4, Mamba) → Video, motion and generative Mamba models

Abstract (arXiv)

Spatio-temporal video prediction plays a pivotal role in critical domains, ranging from weather forecasting to industrial automation. However, in high-precision industrial scenarios such as semiconductor manufacturing, the absence of specialized benchmark datasets severely hampers research on modeling and predicting complex processes. To address this challenge, we make a twofold contribution.First, we construct and release the Chip Dicing Lane Dataset (CHDL), the first public temporal image dataset dedicated to the semiconductor wafer dicing process. Captured via an industrial-grade vision system, CHDL provides a much-needed and challenging benchmark for high-fidelity process modeling, defect detection, and digital twin development.Second, we propose DIFFUMA, an innovative dual-path prediction architecture specifically designed for such fine-grained dynamics. The model captures global long-range temporal context through a parallel Mamba module, while simultaneously leveraging a diffusion module, guided by temporal features, to restore and enhance fine-grained spatial details, effectively combating feature degradation. Experiments demonstrate that on our CHDL benchmark, DIFFUMA significantly outperforms existing methods, reducing the Mean Squared Error (MSE) by 39% and improving the Structural Similarity (SSIM) from 0.926 to a near-perfect 0.988. This superior performance also generalizes to natural phenomena datasets. Our work not only delivers a new state-of-the-art (SOTA) model but, more importantly, provides the community with an invaluable data resource to drive future research in industrial AI.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Video, motion and generative Mamba models, one of 10 directions in State-space sequence models (S4, Mamba). Builds on 31 of its 46 references that are in the map. First posted on arXiv 9 Jul 2025.

Builds on

31 of 46 references.

References · 31

  1. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  2. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
  3. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,403
  4. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,058
  5. 2023Mamba: Linear-Time Sequence Modeling with Selective State Spacescited by 4,718
  6. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,444
  7. 2014Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translationcited by 4,057
  8. 2022Elucidating the Design Space of Diffusion-Based Generative Modelscited by 3,045
  9. 2015Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcastingcited by 1,436
  10. 2017Spatio-Temporal Graph Convolutional Networks: A Deep Learning Framework for Traffic Forecastingcited by 1,136
  11. 2019Graph WaveNet for Deep Spatial-Temporal Graph Modelingcited by 829
  12. 2022Simplified State Space Layers for Sequence Modelingcited by 788
19 more
  1. 2021Combining Recurrent, Convolutional, and Continuous-time Models with Linear State-Space Layerscited by 608
  2. 2021CSDI: Conditional Score-based Diffusion Models for Probabilistic Time Series Imputationcited by 582
  3. 2020Adaptive Graph Convolutional Recurrent Network for Traffic Forecastingcited by 530
  4. 2015Unsupervised Learning of Video Representations using LSTMscited by 435
  5. 2018T-GCN: A Temporal Graph ConvolutionalNetwork for Traffic Predictioncited by 401
  6. 2016Structured Sequence Modeling with Graph Convolutional Recurrent Networkscited by 275
  7. 2020WeatherBench: A benchmark dataset for data-driven weather forecastingcited by 275
  8. 2022Earthformer: Exploring Space-Time Transformers for Earth System Forecastingcited by 152
  9. 2023Spatio-Temporal Graph Neural Networks for Predictive Learning in Urban Computing: A Surveycited by 140
  10. 2018PredRNN++: Towards A Resolution of the Deep-in-Time Dilemma in Spatiotemporal Predictive Learningcited by 122
  11. 2022Flowformer: Linearizing Transformers with Conservation Flowscited by 69
  12. 2024A Survey on Diffusion Models for Time Series and Spatio-Temporal Datacited by 67
  13. 2022Transformers in Remote Sensing: A Surveycited by 61
  14. 2023PastNet: Introducing Physical Inductive Biases for Spatio-temporal Video Predictioncited by 44
  15. 2024Graph Transformers: A Surveycited by 32
  16. 2023Graph Neural Processes for Spatio-Temporal Extrapolationcited by 19
  17. 2025Turb-L1: Achieving Long-term Turbulence Tracing By Tackling Spectral Biascited by 9
  18. 2025Advanced Long-term Earth System Forecastingcited by 9
  19. 2022Physics-Informed Graph Neural Network for Spatial-temporal Production Forecastingcited by 8

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Diff4Splat: Controllable 4D Scene Generation with Latent Dynamic Reconstruction Models
  2. 2026Predict to Skip: Linear Multistep Feature Forecasting for Efficient Diffusion Transformers
  3. 2025Epona: Autoregressive Diffusion World Model for Autonomous Driving
  4. 2024DDPM-MoCo: Advancing Industrial Surface Defect Generation and Detection with Generative and Contrastive Learning
  5. 2024DefFiller: Mask-Conditioned Diffusion for Salient Steel Surface Defect Generation
  6. 2025Noise Fusion-based Distillation Learning for Anomaly Detection in Complex Industrial Environments
  7. 2025DiffuEraser: A Diffusion Model for Video Inpainting
  8. 2025LongDWM: Cross-Granularity Distillation for Building a Long-Term Driving World Model