Learning Video Dynamics with Predictive Differentiable Rendering
About
How to accurately predict a high-fidelity future world? While the visual world is inherently continuous, existing deterministic video prediction models operate in discrete pixel space and are mainly optimized with pixel-wise mean squared error (MSE), which often leads to over-smoothed predictions and a lack of fine-grained visual details. To address these limitations, we propose Predictive Differentiable Rendering (PDR), a novel end-to-end video prediction paradigm that bridges the gap between discrete and continuous representations. Inspired by recent progress in 3D reconstruction with 3D Gaussian Splatting, we introduce PredGS, a lightweight and plug-and-play adapter based on 2D Gaussian representation, which could be seamlessly integrated with existing pixel space predictors, significantly improving spatial detail preservation with negligible computational overhead. Furthermore, we develop predgsplat, a CUDA-accelerated differentiable 2D Gaussian renderer supporting arbitrary channels. Each Gaussian is defined by 5 + C learnable parameters (position, scale, rotation, and C channel amplitudes) and achieves up to 10x faster rendering than the baseline. Optimized by a combined L1 and SSIM loss, PDR overcomes the inherent blurring tendencies of MSE Loss, significantly enhancing the prediction performance. Extensive experiments on diverse real-world benchmarks, including TaxiBJ, WeatherBench, KTH, and Human3.6M, demonstrate that PDR consistently surpasses existing methods, delivering superior detail preservation, visual fidelity, and predictive accuracy.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Video Prediction | KTH 10 -> 20 steps (test) | PSNR26.98 | 102 | |
| Video Prediction | Human3.6M 4 → 4 frames (test) | SSIM98.57 | 16 | |
| Video Prediction | TaxiBJ 4 → 4 frames (test) | MSE0.2807 | 15 | |
| Video Prediction | WeatherBench T2m 12 -> 12 frames (test) | MSE1.071 | 14 | |
| Video Prediction | KTH 64x64, 10 → 30 frames | PSNR26.69 | 7 |