STRIDE: Scene Transition via Interpolative Diffusion Estimation

Video transition generation, the synthesis of coherent intermediate frames bridging a given start and end frame, is a core problem in generative video modeling. Existing diffusion and flow methods are not only computationally heavy but also rely on large pretrained image/video backbones. We present STRIDE ($\text{S}$cene $\text{Tr}$ansition via $\text{I}$nterpolative $\text{D}$iffusion $\text{E}$stimation), an inference-efficient bidirectional latent-diffusion framework whose 3D U-Net denoiser ($\sim$1.54B parameters) is customized and trained as generative-prior-free, that carries no transition or motion prior from any pretrained generator (a frozen VAE and CLIP serve only as fixed interfaces). STRIDE couples position-aware progressive masking, coupled bidirectional temporal attention with a forward/backward consistency loss, and multi-anchor CLIP conditioning, is trained under a four-phase motion-filtered curriculum with an annealed-fusion inference schedule. Operating in a compact $256^2$ pixel and $32^2$ latent space, it infers $[3.7$ to $43]\times$ faster than DynamiCrafter, SEINE, TVG, and Wan2.1-VACE-1.3B, at $\approx [36$ to $181]\times$ lower per-clip compute than DynamiCrafter, SEINE, and TVG. Further, it attains the best inter-frame structural consistency (VSI, FSIM, and tMS-SSIM on all three benchmark datasets i.e., Morph-Bench, TC-Bench, and TransitBench and tPSNR on one of the three). Although SVD-KFI (Generative Inbetweening) wins the four perceptual/appearance-similarity metrics (CLIPSim, DreamSIM, LPIPS, EMD) on all three benchmarks, its inference is $\approx 49\times$ slower than STRIDE, with Wan and Framer occupying intermediate positions. Because all eight metrics are computed between consecutive frames, they measure inter-frame smoothness rather than motion fidelity. This makes the resulting frontier an operating-point trade-off between inter-frame structural consistency and perceptual/appearance similarity, rather than a quality ranking. In our human study (5 models $\times$ 48 clips $\times$ 3 raters $=720$ ratings), we observe high inter-rater agreement (Krippendorff's $\alpha =0.84$–$0.87$). The Friedman omnibus test reveals significant differences in transition-smoothness ratings across models ($p<10^{-14}$). Furthermore, Holm-corrected pairwise Wilcoxon tests confirm that STRIDE is significantly preferred over all four pretrained-prior baselines (DynamiCrafter, SEINE, TVG, Wan2.1-VACE-1.3B). Despite using no pretrained diffusion backbone, STRIDE achieves significantly smoother transitions than pretrained-prior baselines, at the cost of occasional reductions in intermediate-frame visual quality near the transition midpoint.