arXiv Computer Vision

TT-VidT: Decoupling the Temporal Axis for Efficient Motion-Centric Video Pretraining

TT-VidT is a video pretraining method that decouples the temporal axis by combining a per‑frame ViT-B/16 spatial encoder with a compact Temporal Transfer Layer trained via Diff Compression. The authors conduct a systematic 24‑configuration study to isolate architecture, objective, and decoder effects, showing that the full TT-VidT design yields the strongest motion‑sensitive representations. In downstream fine‑tuning, TT‑VidT outperforms state‑of‑the‑art baselines on Jester, Something‑Something V2, ARID, and Diving48 while using significantly fewer encoder FLOPs.

arXiv AI
Aug 28

LeVJEPA: Efficient & Scalable Video Pretraining without the Heuristics

LeVJEPA is a video encoder that eliminates the need for architectural asymmetries, exponential-moving-average target encoders, stop-gradients, and capacity-limited predictors used in prior self‑supervised methods. It trains a single encoder with an invariance loss over global and local views, regularized by SIGReg to prevent collapse, and achieves strong performance with far less pretraining compute. The approach also allows block‑causal attention, making temporal ordering a property of the encoder itself, and matches or surpasses state‑of‑the‑art baselines on both appearance‑centric and motion‑centric benchmarks.

By Lukas Kuhn, Lucas Maes, Giuseppe Serra, Quentin Le Lidec, Yann LeCun, Randall Balestriero, Florian Buettner
arXiv Computer Vision
Aug 27

LongVU-TTT: Causal Test-Time Training for Visual Resampling in Long Video Understanding

LongVU‑TTT is a causal test‑time training method for long‑video multimodal large language models that inserts a convolutional resampler with fast‑weight updates between the vision encoder and the LLM. The fast weights adapt per video and contextualize frame features before compression, while a hybrid selector keeps explicit visual evidence for downstream reasoning. Experiments show that TTT‑Conv outperforms TTT‑MLP and bidirectional Mamba2 on MLVU, and beats attention‑ and fixed‑state recurrent resamplers on three benchmarks, achieving competitive results on five video‑understanding tasks after reducing 512 frames to 128 LLM frames.

By Mahmoud Ahmed, Sameh Abdulah, Olatunji Ruwase, Sam Ade Jacobs, Mathis Bode, Mohamed Elhoseiny
Hugging Face Trending Papers
Aug 13

V-RAE: Rethinking Video Latent Spaces for Generation

Latent video generation relies on autoencoders to define a compact space in which generative models operate. Although video autoencoder architectures have evolved substantially, their latent spaces are still optimized primarily for pixel-level reconstruction and provide limited high-level semantic organization.

arXiv Computer Vision
Oct 22

Rethinking Video Autoencoders for Interactive World Models

The paper introduces CADRE, a family of causal diffusion representation autoencoders designed for video world models that balance spatio‑temporal compression. By training the same world model across different CADRE variants in manipulation and navigation tasks, the authors map the latency‑quality trade‑off and uncover that strong compression favors diffusion decoders, temporal compression is more effective in the autoencoder for manipulation, and spatial compression is preferable when the camera moves. The study also reveals that coarse spatial tokens degrade motion perception more than image quality when the camera is static.

By Adrien Ramanana Rahary, Yutong Bai, Nicolas Dufour, Patrick P\'erez, David Picard
arXiv Computer Vision
Nov 5

CoDecide: Content-Adaptive Video Decoding for High-Throughput Video Analytics

CoDecide is a retrospective analytics system that selectively decodes video frames only when new content appears, using compressed‑bitstream signals to identify novelty. It trains a lightweight compressed‑domain student model with a label‑free novelty signal from a frozen DINOv2 teacher, enabling it to generalize across camera deployments without per‑camera retraining. In experiments on VIRAT, MEVA, and ShanghaiTech, CoDecide outperforms prior compressed‑domain selectors and a dense per‑frame detector, achieving 15 to 24× faster processing while recovering more moving‑object detections under the same decoding budget.

By Shristy Shah, Thomas B. Moeslund, Kamal Nasrollahi
arXiv AI
Sep 28

GVCC: Zero-Shot Video Compression via Codebook-Driven Stochastic Rectified Flow

The paper introduces GVCC, a zero‑shot video compression framework that uses a pretrained generative video model as the decoder. GVCC transforms deterministic rectified‑flow samplers into stochastic processes, enabling the transmission of compressed information through per‑step stochastic innovations. The authors evaluate three GVCC variants—Text‑to‑Video, Image‑to‑Video, and First‑Last‑Frame‑to‑Video—on the UVG dataset, reporting perceptual, fidelity, and temporal metrics without claiming global rate‑distortion gains.

By Ziyue Zeng, Xun Su, Haoyuan Liu, Bingyu Lu, Yui Tatsumi, Hiroshi Watanabe
arXiv AI
Sep 4

LRConv-NeRV: Low Rank Convolution for Efficient Neural Video Compression

LRConv-NeRV introduces low‑rank separable convolutions into the NeRV neural video decoder, replacing selected dense 3x3 layers to reduce computational load and memory usage. By applying low‑rank factorization progressively from the largest to earlier decoder stages, the method offers controllable trade‑offs between reconstruction quality and efficiency. Experiments show that applying LRConv only to the final decoder stage cuts decoder complexity by 68% and model size by 9.3% with negligible quality loss, while INT8 quantization preserves performance close to the dense baseline.

By Tamer Shanableh