GazeFlow is a new framework for egocentric gaze prediction that models gaze as a joint distribution of temporal positions conditioned on both top‑down task cues and bottom‑up visual saliency. It employs conditional flow matching to iteratively transform Gaussian noise into realistic gaze trajectories, using a velocity field informed by video‑encoded visual features and global task queries. On standard benchmarks, GazeFlow outperforms existing methods on per‑frame metrics and produces trajectories that better reflect human gaze dynamics.
By Sheng Zhao, Weikai Lin, Yuhao Zhu
arXiv:2610.28502v1 Announce Type: new
Abstract: Smart glasses can help users understand their activities and surroundings, but small vision-language models must preserve useful evidence within limite...
By Trong-Thang Pham, Anh Nguyen, Hien Van Nguyen, Ngan Le
arXiv:2610.00922v1 Announce Type: new
Abstract: Gaze estimation under natural head-eye motion underpins applications from driver monitoring to human-computer interaction. Single-frame methods predict...
By Jungmin Lee, Niamat Ullah, Yoseob Han
The paper introduces the Causal Context-Gated Forecaster (CCGF) for predicting a driver's gaze during dashboard-mounted tracker dropouts. CCGF uses a 60‑frame history of gaze and head pose combined with DINOv3 scene features, and a learned reliability gate adjusts the influence of these inputs as the dropout progresses. Experiments on 2,047 naturalistic driving events show that live scene updates reduce median error by 33% compared to history‑only forecasting, while frozen scene input yields higher error, demonstrating the value of real‑time scene information.
By Shabnam Shabani, Ghazal Farhani
arXiv:2608. 08947v1 Announce Type: cross Abstract: Current hazard detection systems in autonomous driving may develop mesa objectives, learned internal goals that achieve high training performance through spurious correlations rather than genuine hazard recognition.
By Lennox Anderson, Ahmed Boutar, Jonah Mulcrone, Tal Erez
The paper introduces SalTempto, a new video saliency benchmark featuring 224 highly dynamic clips from HACS‑Segments, each with event context and gaze data from up to 16 subjects. It shows that static baselines perform poorly on SalTempto compared to LEDOV, while fine‑tuned temporal models achieve significant gains, yet still leave a large portion of explainable gaze information unaccounted for. The dataset highlights the need for better temporal modeling in video saliency research.
By Susmit Agrawal, Rebecca Wanner, Juliane Verwiebe, Matthias Tangemann, Matthias Bethge, Matthias K\"ummerer
GazeFS is a model that predicts and stabilizes target‑centered gaze trajectories using a variable‑length gaze‑head history, without requiring target information during inference. It maps this history to the next target‑center direction and a short‑horizon Search/Focus estimate, improving focus target centering and reducing residual gaze error. Across 7,960 acquisition episodes from 30 participants, GazeFS reduces Focus episode bias, dispersion, and P90 target error by 0.182°, 0.257°, and 0.400°, respectively, while maintaining high phase‑balanced accuracy and AUPRC.
By Yaozheng Xia, Zaiping Zhu, Bo Pang, Minghao Xie, Hui Li, Shaorong Wang, Sheng Li
arXiv:2608.22926v1 Announce Type: new
Abstract: Gaze is increasingly used as an input signal for vision and multimodal models, yet no consensus exists on how to represent it across datasets. Raw trac...
By Virmarie Maquiling, Zhuojiang Cai, Enkelejda Kasneci
Vision-language models excel at video captioning, yet typically generate descriptions that fail to capture individual viewers' attention. We propose VEGAS (Video caption Evaluation via GAze Score), a training-free metric that leverages test-time gaze to sample personalized, attention-aligned text.
arXiv:2602. 14834v2 Announce Type: replace-cross Abstract: Human eye movements in visual recognition reflect a balance between foveal sampling and peripheral context.
By Pengcheng Pan, Yonekura Shogo, Yasuo Kuniyosh
Gaze Attention is a new mechanism for multimodal large language models that selectively focuses on relevant visual regions during each generation step, rather than attending to all visual tokens. By grouping tokens into spatial regions and using learnable context tokens to retain global information, it reduces attention computation and visual key‑value entries by up to 90%. Experiments on 13 image and 6 video benchmarks show that Gaze Attention matches or outperforms dense‑attention baselines while using fewer visual resources.
By Junha Song, Byeongho Heo, Geonmo Gu, Jaegul Choo, Dongyoon Han, Sangdoo Yun
arXiv:2607. 08489v1 Announce Type: cross Abstract: Vision-language models excel at video captioning, yet typically generate descriptions that fail to capture individual viewers' attention.
By Shenghui Chen, Po-han Li, Ximeng Sun, Shijia Yang, Emad Barsoum, Zicheng Liu, Sandeep Chinchali, Ufuk Topcu