LDDR (Linear DPP Dynamic Resolution) is a training‑free, budget‑aware video frame sampling framework designed for multimodal large language models (MLLMs). It uses query‑aware Determinantal Point Process (DPP) selection in a task‑conditioned feature space, achieving a 3× runtime speedup over standard DPP baselines. By assigning a non‑uniform token budget—allocating more tokens to informative, non‑redundant frames and downscaling or pruning less useful ones—LDDR improves performance by 2.5 points in low‑budget settings and 1.6 points in high‑budget scenarios across four video benchmarks and multiple MLLM families.
By Jingfeng Chen, Jiawen Qian, Wendi Deng, Yinuo Guo, Jiaqi Yu, Sicong Leng, Raghuveer Thirukovalluru, Bhuwan Dhingra
arXiv:2606.06158v2 Announce Type: replace
Abstract: Adaptive video tokenisation seeks to dynamically allocate token budgets based on the underlying visual complexity of a sequence. Current continuous...
By Kevin Dave, Sai Aditya Patkuri, Chhaya Kumar Das, Gouranga Bala, Rajeshkumar SA, R. Venkatesh Babu
arXiv:2608.20473v1 Announce Type: new
Abstract: Video language models process videos as dense visual-token sequences with substantial representational redundancy. Compressing these sequences is there...
By Wenti Yin, Xiaotian Han, Junyuan Shang, Yuchen Ding, Shuohuan Wang, Dianhai Yu, Changxin Gao, Nong Sang
arXiv:2605.16981v3 Announce Type: replace
Abstract: Streaming 3D reconstruction under a strict constant-memory budget hinges on how the recurrent state is updated as the stream evolves. We profile TT...
By Kejun Ren, Lei Jin, Tianxin Huang, Lianming Xu, Li Wang
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
The paper introduces UniGVC, a unified generative video codec that supports both low‑delay streaming and high‑quality playback from a single bitstream. It offers two decoding modes: UniGVC‑BI, a bidirectional diffusion model with dense attention for quality‑oriented random‑access decoding, and UniGVC‑AR, a windowed autoregressive diffusion model for low‑delay streaming. Both modes share a hierarchical latent codec that produces ultra‑low‑bitrate conditions, enabling the BI model to act as a teacher for distilling the AR model, while a lightweight conditional decoder further reduces streaming latency.
By Tianyu Zhang, Shuai Huo, Dengchao Jin, Hewei Liu, Meng Lei, Tianwen Wu, Jiawen Gu, Chao Zhou, Li Li, Dong Liu
The paper introduces a zero‑shot subject‑driven video generation framework that eliminates the need for per‑subject tuning and large subject‑video datasets. It achieves this by separating identity injection—learned from subject‑image pairs—and motion‑awareness preservation—maintained with a small set of arbitrary videos, and optimizes both with stochastic switching and dropout techniques. Using CogVideoX‑5B, the method adapts a single model with only 200K subject‑image pairs and 4,000 arbitrary videos in 288 A100 GPU hours, representing roughly 1% of the compute required by previous zero‑shot baselines while preserving subject fidelity and motion quality.
By Daneul Kim, Jingxu Zhang, Wonjoon Jin, Sunghyun Cho, Qi Dai, Jaesik Park, Chong Luo
arXiv:2607. 24570v1 Announce Type: cross Abstract: Large-scale video platforms process millions of uploads hourly, requiring moderation systems that can localize when and where policy violations occur within each video.
By Jiameng Zhang, Srikanth Madikeri
The paper introduces a resolution‑flexible decoder for hybrid neural video representations, using uniform 2× upsampling stages that adapt to any target resolution by tracing spatial sizes backward from the final output. After each stage, feature maps are aligned to the desired size via minimal padding or cropping, and the process is guided by intermediate reconstruction supervision and a difficulty‑aware frame sampling strategy. Experiments on the UVG dataset demonstrate that this framework improves reconstruction quality compared to existing NVR baselines.
By Taiga Hayami, Masaya Takabe, Hiroshi Watanabe
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.
By Shih-Ying Yeh, Daniel Z. Kaplan, Xuehai Wang, Fu-En Yang, Min-Hung Chen, Shang-Hong Lai
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
ShallowStream is a framework for streaming video understanding that uses the shallow layers of a multimodal large language model (MLLM) to encode frames and build a lightweight index. During streaming, it maintains an always‑on index via the KV cache of shallow layers, and at query time it scores context frames using shallow‑layer attention and selects diverse evidence for answering. The approach matches the performance of leading streaming methods while cutting per‑frame prefill latency and 10‑second end‑to‑end latency by up to 52.1× and 11.9×, respectively.
By Jitai Hao, Ke Yang, Qiang Huang, Jun Yu