Saved in:
Bibliographic Details
Main Authors: Chen, Zhan, Guo, Zile, Zhu, Enze, Zhang, Peirong, Liu, Xiaoxuan, Wang, Lei, Zhang, Yidan
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.21710
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911342517026816
author Chen, Zhan
Guo, Zile
Zhu, Enze
Zhang, Peirong
Liu, Xiaoxuan
Wang, Lei
Zhang, Yidan
author_facet Chen, Zhan
Guo, Zile
Zhu, Enze
Zhang, Peirong
Liu, Xiaoxuan
Wang, Lei
Zhang, Yidan
contents Video prediction is plagued by a fundamental trilemma: achieving high-resolution and perceptual quality typically comes at the cost of real-time speed, hindering its use in latency-critical applications. This challenge is most acute for autonomous UAVs in dense urban environments, where foreseeing events from high-resolution imagery is non-negotiable for safety. Existing methods, reliant on iterative generation (diffusion, autoregressive models) or quadratic-complexity attention, fail to meet these stringent demands on edge hardware. To break this long-standing trade-off, we introduce RAPTOR, a video prediction architecture that achieves real-time, high-resolution performance. RAPTOR's single-pass design avoids the error accumulation and latency of iterative approaches. Its core innovation is Efficient Video Attention (EVA), a novel translator module that factorizes spatiotemporal modeling. Instead of processing flattened spacetime tokens with $O((ST)^2)$ or $O(ST)$ complexity, EVA alternates operations along the spatial (S) and temporal (T) axes. This factorization reduces the time complexity to $O(S + T)$ and memory complexity to $O(max(S, T))$, enabling global context modeling at $512^2$ resolution and beyond, operating directly on dense feature maps with a patch-free design. Complementing this architecture is a 3-stage training curriculum that progressively refines predictions from coarse structure to sharp, temporally coherent details. Experiments show RAPTOR is the first predictor to exceed 30 FPS on a Jetson AGX Orin for $512^2$ video, setting a new state-of-the-art on UAVid, KTH, and a custom high-resolution dataset in PSNR, SSIM, and LPIPS. Critically, RAPTOR boosts the mission success rate in a real-world UAV navigation task by 18%, paving the way for safer and more anticipatory embodied agents.
format Preprint
id arxiv_https___arxiv_org_abs_2512_21710
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle RAPTOR: Real-Time High-Resolution UAV Video Prediction with Efficient Video Attention
Chen, Zhan
Guo, Zile
Zhu, Enze
Zhang, Peirong
Liu, Xiaoxuan
Wang, Lei
Zhang, Yidan
Computer Vision and Pattern Recognition
Video prediction is plagued by a fundamental trilemma: achieving high-resolution and perceptual quality typically comes at the cost of real-time speed, hindering its use in latency-critical applications. This challenge is most acute for autonomous UAVs in dense urban environments, where foreseeing events from high-resolution imagery is non-negotiable for safety. Existing methods, reliant on iterative generation (diffusion, autoregressive models) or quadratic-complexity attention, fail to meet these stringent demands on edge hardware. To break this long-standing trade-off, we introduce RAPTOR, a video prediction architecture that achieves real-time, high-resolution performance. RAPTOR's single-pass design avoids the error accumulation and latency of iterative approaches. Its core innovation is Efficient Video Attention (EVA), a novel translator module that factorizes spatiotemporal modeling. Instead of processing flattened spacetime tokens with $O((ST)^2)$ or $O(ST)$ complexity, EVA alternates operations along the spatial (S) and temporal (T) axes. This factorization reduces the time complexity to $O(S + T)$ and memory complexity to $O(max(S, T))$, enabling global context modeling at $512^2$ resolution and beyond, operating directly on dense feature maps with a patch-free design. Complementing this architecture is a 3-stage training curriculum that progressively refines predictions from coarse structure to sharp, temporally coherent details. Experiments show RAPTOR is the first predictor to exceed 30 FPS on a Jetson AGX Orin for $512^2$ video, setting a new state-of-the-art on UAVid, KTH, and a custom high-resolution dataset in PSNR, SSIM, and LPIPS. Critically, RAPTOR boosts the mission success rate in a real-world UAV navigation task by 18%, paving the way for safer and more anticipatory embodied agents.
title RAPTOR: Real-Time High-Resolution UAV Video Prediction with Efficient Video Attention
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.21710