VisReflect: Latent Visual Reflection for Fine-Grained Perception in Long Visual Context

• URL: http://arxiv.org/abs/2606.30288v1
• Date: Mon, 29 Jun 2026 13:30:17 GMT
• Title: VisReflect: Latent Visual Reflection for Fine-Grained Perception in Long Visual Context
• Abstract summary: We propose **VisReflect**, a framework that improves fine-grained perception in long visual contexts through latent visual reflection.<n>Our method consistently improves over strong baselines, achieving gains of 4.1% on image benchmarks and 1.8% on video benchmarks.<n>Compared with zooming-based methods, our model achieves comparable performance while reducing inference time by roughly 44% on video understanding.
• Score: 47.466197471764104
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large Vision Language Models (LVLMs) have achieved remarkable success on vision-language tasks, yet fine-grained perception over high-resolution images and long-context videos remains challenging. As the number of visual tokens increases, the visual attention sink phenomenon becomes increasingly severe, causing irrelevant tokens to absorb a disproportionate amount of attention mass. Recent approaches attempt to mitigate this issue by explicitly predicting bounding boxes or temporal spans and re-encoding the cropped visual regions. Such methods depend on unreliable numeric localization in the discrete token space and incur significant computational overhead due to additional forward passes. In this work, we propose **VisReflect**, a simple yet effective framework that improves fine-grained perception in long visual contexts through latent visual reflection. Instead of decoding intermediate predictions into discrete tokens, the model generates continuous visual reflection that represents question-relevant visual features in the latent space. These reflections selectively emphasize salient regions or frames, guiding attention towards relevant visual tokens within a single forward pass. We conduct comprehensive evaluations on challenging high-resolution image benchmarks, including BLINK, V*, and HRBench-4K/8K, as well as video understanding benchmarks such as MVBench, VideoMME, and MLVU. Our method consistently improves over strong baselines, achieving gains of 4.1% on image benchmarks and 1.8% on video benchmarks. Compared with zooming-based methods, our model achieves comparable performance while reducing inference time by roughly 44% on video understanding.

Related papers

• LASER: A Corrective Lens for LVLMs via Visual Attention Preservation and Sink Suppression [50.454633334898396]
  Large vision-language models (LVLMs) exhibit strong reasoning ability but suffer from visual forgetting during long-horizon decoding.<n>We propose LASER, a post-training framework that regulates both the visual attention trajectory and intra-visual token attention distribution during reasoning.<n>Experiments on eight benchmark datasets demonstrate that LASER consistently outperforms strong baselines.
  arXiv  Detail & Related papers  (2026-07-02T04:59:56Z)
• Learning to See What You Need: Gaze Attention for Multimodal Large Language Models [96.20985292033465]
  We introduce Gaze Attention, a novel mechanism that enables MLLMs to selectively attend to task-relevant visual regions during generation.<n>We show that Gaze Attention matches or surpasses dense-attention baselines, while using up to 90% fewer visual KV entries in the attention computation.
  arXiv  Detail & Related papers  (2026-05-13T06:54:09Z)
• Sparrow: Text-Anchored Window Attention with Visual-Semantic Glimpsing for Speculative Decoding in Video LLMs [28.766303423132722]
  Video Large Language Models (Vid-LLMs) typically fall into the trap of attention dilution and negative visual gain due to key-value cache explosion and context window mismatches.<n>We propose the Sparrow framework, which first utilizes visually-aware text-anchored window attention via hidden state reuse to fully offload visual computation to the target model.<n>Experiments show that Sparrow achieves an average speedup of 2.82x even with 25k visual tokens, effectively resolving the performance degradation in long sequences.
  arXiv  Detail & Related papers  (2026-02-17T02:51:36Z)
• VideoAnchor: Reinforcing Subspace-Structured Visual Cues for Coherent Visual-Spatial Reasoning [69.64660280965971]
  VideoAnchor is a plug-and-play module that leverages subspace affinities to reinforce visual cues across frames without retraining.<n>We show consistent performance gains on benchmarks with InternVL2-8B and Q2.5VL-72B.<n>Our codes will be made public at https://github.com/feufhd/VideoAnchor.
  arXiv  Detail & Related papers  (2025-09-29T17:54:04Z)
• Autoregressive Semantic Visual Reconstruction Helps VLMs Understand Better [44.15671594378141]
  We introduce Autoregressive Semantic Visual Reconstruction (ASVR), which enables joint learning of visual and textual modalities within a unified autoregressive framework.<n>ASVR improves LLaVA-1.5 by 5% in average scores across 14 multimodal benchmarks.
  arXiv  Detail & Related papers  (2025-06-10T17:57:50Z)
• TokBench: Evaluating Your Visual Tokenizer before Visual Generation [75.38270351179018]
  We analyze text and face reconstruction quality across various scales for different image tokenizers and VAEs.<n>Our results show modern visual tokenizers still struggle to preserve fine-grained features, especially at smaller scales.
  arXiv  Detail & Related papers  (2025-05-23T17:52:16Z)
• MINT: Mitigating Hallucinations in Large Vision-Language Models via Token Reduction [6.416957959150438]
  Hallucinations hinder the application of Large Vision-Language Models (LVLMs) in domains that require high reliability.<n>We propose MINT, a training-free decoding strategy, MItigating hallucinations via tokeN reducTion.<n>Our approach achieves a 4% improvement in mitigating hallucinations caused by distracted perception compared to original models.
  arXiv  Detail & Related papers  (2025-02-02T08:34:57Z)
• VideoLLM-MoD: Efficient Video-Language Streaming with Mixture-of-Depths Vision Computation [66.00245701441547]
  We introduce a novel approach to reduce vision compute by leveraging redundant vision tokens "skipping layers" rather than decreasing the number of vision tokens. Our method, VideoLLM-MoD, is inspired by mixture-of-depths LLMs and addresses the challenge of numerous vision tokens in long-term or streaming video.
  arXiv  Detail & Related papers  (2024-08-29T17:21:58Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.