Dual-Stream Cross-Anchor Correction Grounding Long-Form Captions and the Domain Limits of Object-Level Anchors

• URL: http://arxiv.org/abs/2608.12746v3
• Date: Wed, 19 Aug 2026 06:13:49 GMT
• Title: Dual-Stream Cross-Anchor Correction Grounding Long-Form Captions and the Domain Limits of Object-Level Anchors
• Abstract summary: This paper proposes Dual-Stream Cross-Anchor Correction (DSCC)<n>Unlike work that post-processes decoding, DSCC injects object-level visual anchors into the language model itself during fine-tuning.<n> experiments span long-caption hallucination, object-existence discrimination and cross-domain generalisation.
• Score: 7.647944993041615
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Object hallucination in multimodal large language models arises when language priors and corpus co-occurrence bias outweigh the visual evidence, with nothing tying an object mention to the image. Most remedies intervene at decoding time, yet under a unified protocol their benefit is confined to short captions; supervised fine-tuning (SFT) on a detail-rich corpus lengthens captions, but over forty percent still name absent objects. This paper proposes Dual-Stream Cross-Anchor Correction (DSCC). Unlike work that post-processes decoding, DSCC injects object-level visual anchors into the language model itself during fine-tuning: a perception stream aligns object-level hidden states at an intermediate layer to frozen text anchors by a bidirectional contrastive objective; a cognition stream lets deeper layers query those anchors by cross-attention at every generation step; and a two-stage curriculum gate couples them, making evidence retrieval a structural constraint on generation. Under one backbone and one scoring protocol, experiments span long-caption hallucination, object-existence discrimination and cross-domain generalisation, with vanilla SFT on the same corpus and schedule as a length- and density-matched control separating the data effect from the architectural gain. DSCC alone reaches the long-caption, low-hallucination region: captions roughly 1.9 times the baseline length at 88.19% precision per object mention, the highest under a density-independent criterion. Ablations expose a synergy: the perception stream alone degrades precision yet reverses sign when stacked on the cognition stream. No universal superiority is claimed: three out-of-domain benchmarks yield a predictable, falsifiable domain-conditionality, the synergy being bound to the anchors' semantic domain and breaking on charts and illusions.

Related papers

• VPRef: A Cross-Domain Benchmark for Referring Remote Sensing Image Segmentation [17.658559673931965]
  This paper establishes the first cross-domain RRSIS benchmark, comprising 46,972 language-image-annotation triplets.<n>We develop a tailored parameter-efficient domain adaptation baseline anchored on the Segment Anything Model (SAM3) via Low-Rank Adaptation (LoRA)<n>Our framework counteracts visual distribution discrepancies through pseudo-label-driven self-training and addresses textual logic drift via random multi-granularity text prompt mixing.
  arXiv  Detail & Related papers  (2026-09-15T01:18:06Z)
• ENEAS: Embedding-guided Neural Ensemble for Adaptive Segmentation [36.94429692322632]
  We present ENEAS, a text-promptable method for instance tracking and semantic discovery.<n>We extend the geometrically robust SeC architecture, previously limited to point interactions, with a text-prompting adapter.<n>The verification layer combines high-speed visual embedding matching with conditional VLM refinement, invoking semantic reasoning only for ambiguous candidates.
  arXiv  Detail & Related papers  (2026-09-03T12:25:26Z)
• A Dynamic-Semantics Framework for Grounding Human Referring Expressions in Visual Perceptual Data [0.0]
  We present a framework that externalizes pact state into three explicit, inspectable sets of referent-object bindings.<n>We evaluate the framework on the Stanford Repeated Reference Game corpus.
  arXiv  Detail & Related papers  (2026-08-09T12:14:10Z)
• Beyond Static Anchors: Bounded Prototype Conditioning for Language-Free Medical Anomaly Detection [52.986874008247554]
  We propose ReCAP, a language-free framework that replaces static anchors with input-conditioned visual prototypes.<n>ReCAP re-centers separated normal and abnormal prototypes for each image through a bounded gated modulation.<n>It achieves the best image-level AUROC on all zero-shot and 23 of 24 few-shot settings, and the best zero-shot pixel-level AUROC on all three segmentation datasets.
  arXiv  Detail & Related papers  (2026-08-01T04:46:55Z)
• Hierarchical Semantic-Constrained Heterogeneous Graph for Audio-Visual Event Localization [69.67186845167568]
  Open-vocabulary audio-visual event localization (OV-AVEL) jointly models audio-visual cues to recognize and temporally localize events.<n>Existing methods primarily learn joint audio-visual representations in Euclidean space.<n>We propose a hierarchical semantic constrained heterogeneous graph (HSCHG) for audio-visual event localization framework.
  arXiv  Detail & Related papers  (2026-06-05T08:23:58Z)
• Evading Visual Aphasia: Contrastive Adaptive Semantic Token Pruning for Vision-Language Models [52.78477729846771]
  We introduce COAST (COntrastive Adaptive Semantic Token Pruning), a training-free pruning framework that casts compression as adaptive semantic routing.<n> COAST consistently outperforms strong pruning baselines across token budgets and generalizes across multiple LVLM families.
  arXiv  Detail & Related papers  (2026-05-10T09:07:04Z)
• STAND: Semantic Anchoring Constraint with Dual-Granularity Disambiguation for Remote Sensing Image Change Captioning [23.80607349547519]
  Remote sensing image change captioning (RSICC) aims to describe the difference between two remote sensing images.<n>Recent methods have explored video modeling, but they largely overlook the inherent ambiguities in viewpoint, scale, and prior knowledge.<n>We present STAND, a Semantic Anchoring Constraint with Dual-Granularity Disambiguation for RSICC, to progressively resolve these ambiguities.
  arXiv  Detail & Related papers  (2026-04-25T13:50:11Z)
• Explaining, Verifying, and Aligning Semantic Hierarchies in Vision-Language Model Embeddings [10.299190830046872]
  We present a post-hoc framework to explain, verify, and align semantic hierarchies induced by a VLM over a given set of child classes.<n>We quantify plausibility by comparing the extracted tree against human retrieval using efficient tree-aware and edge-level consistency measures, and we evaluate utility via explainable hierarchical tree-traversal inference.
  arXiv  Detail & Related papers  (2026-03-26T00:38:38Z)
• HGNet: Scalable Foundation Model for Automated Knowledge Graph Generation from Scientific Literature [6.001574550157585]
  We propose a two-stage framework for scalable, zero-shot scientific knowledge graph constructions.<n>We release SPHERE, a benchmark for hierarchical relation extraction.<n>In zero-shot settings, gains reach 10.76% for NER and 26.2% for RE.
  arXiv  Detail & Related papers  (2026-03-24T12:35:10Z)
• Hierarchical Self-Supervised Representation Learning for Depression Detection from Speech [51.14752758616364]
  Speech-based depression detection (SDD) is a promising, non-invasive alternative to traditional clinical assessments.<n>We propose HAREN-CTC, a novel architecture that integrates multi-layer SSL features using cross-attention within a multitask learning framework.<n>The model achieves state-of-the-art macro F1-scores of 0.81 on DAIC-WOZ and 0.82 on MODMA, outperforming prior methods across both evaluation scenarios.
  arXiv  Detail & Related papers  (2025-10-05T09:32:12Z)
• SOOD: Towards Semi-Supervised Oriented Object Detection [57.05141794402972]
  This paper proposes a novel Semi-supervised Oriented Object Detection model, termed SOOD, built upon the mainstream pseudo-labeling framework. Our experiments show that when trained with the two proposed losses, SOOD surpasses the state-of-the-art SSOD methods under various settings on the DOTA-v1.5 benchmark.
  arXiv  Detail & Related papers  (2023-04-10T11:10:42Z)
• Bridging the Gap between Object and Image-level Representations for Open-Vocabulary Detection [54.96069171726668]
  Two popular forms of weak-supervision used in open-vocabulary detection (OVD) include pretrained CLIP model and image-level supervision. We propose to address this problem by performing object-centric alignment of the language embeddings from the CLIP model. We establish a bridge between the above two object-alignment strategies via a novel weight transfer function.
  arXiv  Detail & Related papers  (2022-07-07T17:59:56Z)
• Language-Bridged Spatial-Temporal Interaction for Referring Video Object Segmentation [28.472006665544033]
  Referring video object segmentation aims to predict foreground labels for objects referred by natural language expressions in videos. Previous methods either depend on 3D ConvNets or incorporate additional 2D ConvNets as encoders to extract mixed spatial-temporal features. We propose a Language-Bridged Duplex Transfer (LBDT) module which utilizes language as an intermediary bridge to accomplish explicit and adaptive spatial-temporal interaction earlier in the encoding phase.
  arXiv  Detail & Related papers  (2022-06-08T10:12:53Z)
• CDistNet: Perceiving Multi-Domain Character Distance for Robust Text Recognition [87.3894423816705]
  We propose a novel module called Multi-Domain Character Distance Perception (MDCDP) to establish a visually and semantically related position embedding. MDCDP uses the position embedding to query both visual and semantic features following the cross-attention mechanism. We develop CDistNet that stacks multiple MDCDPs to guide a gradually precise distance modeling.
  arXiv  Detail & Related papers  (2021-11-22T06:27:29Z)

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.