English
Related papers

Related papers: Evict3R: Training-Free Token Eviction for Memory-B…

200 papers

Multimodal large language models (MLLMs) suffer from high computational costs due to excessive visual tokens, particularly in high-resolution and video-based scenarios. Existing token reduction methods typically focus on isolated pipeline…

Computer Vision and Pattern Recognition · Computer Science 2026-03-31 Hanxun Yu , Wentong Li , Xuan Qu , Song Wang , Junbo Chen , Jianke Zhu

Online streaming video understanding requires models to process continuous visual inputs and respond to user queries in real time, where the unbounded stream and unpredictable query timing turn memory management into a central challenge.…

Computer Vision and Pattern Recognition · Computer Science 2026-05-11 Hang Wu , Sherin Mary Mathews , Yujun Cai , Ming-Hsuan Yang , Yiwei Wang

Recent advances in Video Large Language Models (VLLMs) have significantly enhanced their ability to understand video content. Nonetheless, processing long videos remains challenging due to high computational demands and the redundancy…

Computer Vision and Pattern Recognition · Computer Science 2025-09-30 Mengyue Wang , Shuo Chen , Kristian Kersting , Volker Tresp , Yunpu Ma

Vision Transformers (ViTs) have emerged as state-of-the-art models for various vision tasks recently. However, their heavy computation costs remain daunting for resource-limited devices. To address this, researchers have dedicated…

Computer Vision and Pattern Recognition · Computer Science 2025-10-20 Ao Wang , Hui Chen , Zijia Lin , Sicheng Zhao , Jungong Han , Guiguang Ding

We present STream3R, a novel approach to 3D reconstruction that reformulates pointmap prediction as a decoder-only Transformer problem. Existing state-of-the-art methods for multi-view reconstruction either depend on expensive global…

Computer Vision and Pattern Recognition · Computer Science 2025-08-15 Yushi Lan , Yihang Luo , Fangzhou Hong , Shangchen Zhou , Honghua Chen , Zhaoyang Lyu , Shuai Yang , Bo Dai , Chen Change Loy , Xingang Pan

The Visual Geometry Grounded Transformer (VGGT) enables strong feed-forward 3D reconstruction without per-scene optimization. However, its billion-parameter scale creates high memory and compute demands, hindering on-device deployment.…

Hardware Architecture · Computer Science 2026-01-29 Yipu Zhang , Jintao Cheng , Xingyu Liu , Zeyu Li , Carol Jingyi Li , Jin Wu , Lin Jiang , Yuan Xie , Jiang Xu , Wei Zhang

Several video understanding tasks, such as natural language temporal video grounding, temporal activity localization, and audio description generation, require "temporally dense" reasoning over frames sampled at high temporal resolution.…

Computer Vision and Pattern Recognition · Computer Science 2025-09-17 Mattia Soldan , Fabian Caba Heilbron , Bernard Ghanem , Josef Sivic , Bryan Russell

While previous CNN-based models have exhibited promising results for salient object detection (SOD), their ability to explore global long-range dependencies is restricted. Our previous work, the Visual Saliency Transformer (VST), addressed…

Computer Vision and Pattern Recognition · Computer Science 2024-04-12 Nian Liu , Ziyang Luo , Ni Zhang , Junwei Han

Deep learning models are increasingly utilized on resource-constrained edge devices for real-time data analytics. Recently, Vision Transformer and their variants have shown exceptional performance in various computer vision tasks. However,…

Computer Vision and Pattern Recognition · Computer Science 2025-05-22 Xiang Liu , Yijun Song , Xia Li , Yifei Sun , Huiying Lan , Zemin Liu , Linshan Jiang , Jialin Li

Adapting Multimodal Large Language Models (MLLMs) for hour-long videos is bottlenecked by context limits. Dense visual streams saturate token budgets and exacerbate the lost-in-the-middle phenomenon. Existing heuristics, like sparse…

3D Gaussian Splatting (3DGS) has gained popularity for its efficiency and sparse Gaussian-based representation. However, 3DGS struggles to meet the real-time requirement of 90 frames per second (FPS) on resource-constrained mobile devices,…

Graphics · Computer Science 2025-06-12 Chenqi Zhang , Yu Feng , Jieru Zhao , Guangda Liu , Wenchao Ding , Chentao Wu , Minyi Guo

Video-to-video translation aims to generate video frames of a target domain from an input video. Despite its usefulness, the existing networks require enormous computations, necessitating their model compression for wide use. While there…

Computer Vision and Pattern Recognition · Computer Science 2023-10-05 Chaeyeon Chung , Yeojeong Park , Seunghwan Choi , Munkhsoyol Ganbat , Jaegul Choo

Video large language models (Video-LLMs) face high computational costs due to large volumes of visual tokens. Existing token compression methods typically adopt a two-stage spatiotemporal compression strategy, relying on stage-specific…

Computer Vision and Pattern Recognition · Computer Science 2026-03-24 Junhao Du , Jialong Xue , Anqi Li , Jincheng Dai , Guo Lu

Streaming 3D reconstruction under a strict constant-memory budget hinges on how the recurrent state is updated as the stream evolves. We profile TTT3R-style per-token gates across five benchmarks and discover a structural bottleneck: the…

Computer Vision and Pattern Recognition · Computer Science 2026-05-19 Kejun Ren , Lei Jin , Tianxin Huang , Lianming Xu , Li Wang

Vision-Language Models (VLMs) face significant computational challenges in video processing due to massive data redundancy, which creates prohibitively long token sequences. To address this, we introduce Triage, a training-free,…

Computer Vision and Pattern Recognition · Computer Science 2026-02-02 Anmin Wang , Nan Zhang , Wei Tao , Xiaoyang Qu , Guokuan Li , Jiguang Wan , Jianzong Wang

3D Gaussian Splatting (3DGS) has demonstrated remarkable effectiveness in 3D reconstruction, achieving high-quality results with real-time radiance field rendering. However, a key challenge is the substantial storage cost: reconstructing a…

Computer Vision and Pattern Recognition · Computer Science 2025-12-03 Haishan Wang , Mohammad Hassan Vali , Arno Solin

Vision transformers (ViTs) have become the popular structures and outperformed convolutional neural networks (CNNs) on various vision tasks. However, such powerful transformers bring a huge computation burden, because of the exhausting…

Computer Vision and Pattern Recognition · Computer Science 2022-09-13 Zhuofan Zong , Kunchang Li , Guanglu Song , Yali Wang , Yu Qiao , Biao Leng , Yu Liu

Recent research has revealed that reducing the temporal and spatial redundancy are both effective approaches towards efficient video recognition, e.g., allocating the majority of computation to a task-relevant subset of frames or the most…

Computer Vision and Pattern Recognition · Computer Science 2022-09-28 Yulin Wang , Yang Yue , Xinhong Xu , Ali Hassani , Victor Kulikov , Nikita Orlov , Shiji Song , Humphrey Shi , Gao Huang

High-resolution Vision-Language Models (VLMs) are widely used in multimodal tasks to enhance accuracy by preserving detailed image information. However, these models often generate an excessive number of visual tokens due to the need to…

Computer Vision and Pattern Recognition · Computer Science 2024-12-30 Kazi Hasan Ibn Arif , JinYi Yoon , Dimitrios S. Nikolopoulos , Hans Vandierendonck , Deepu John , Bo Ji

The key-value (KV) cache is widely treated as essential state in transformer inference, and a large body of work engineers policies to compress, evict, or approximate its entries. We prove that this state is entirely redundant: keys and…

Machine Learning · Computer Science 2026-03-23 Kaleem Ullah Qasim , Jiashu Zhang , Muhammad Kafeel Shaheen , Razan Alharith , Heying Zhang