A single unified multimodal model matches leading task-specialized vision systems across detection, segmentation, dense geometry, and multi-view 3D by casting all outputs as native text or image generation.
super hub Canonical reference
Shikra: Unleashing Multimodal LLM's Referential Dialogue Magic
Canonical reference. 79% of citing Pith papers cite this work as background.
107
Pith papers citing it
71
external citations · Pith
Background
79% of classified citations
abstract
In human conversations, individuals can indicate relevant regions within a scene while addressing others. In turn, the other person can then respond by referring to specific regions if necessary. This natural referential ability in dialogue remains absent in current Multimodal Large Language Models (MLLMs). To fill this gap, this paper proposes an MLLM called Shikra, which can handle spatial coordinate inputs and outputs in natural language. Its architecture consists of a vision encoder, an alignment layer, and a LLM. It is designed to be straightforward and simple, without the need for extra vocabularies, position encoder, pre-/post-detection modules, or external plug-in models. All inputs and outputs are in natural language form. Referential dialogue is a superset of various vision-language (VL) tasks. Shikra can naturally handle location-related tasks like REC and PointQA, as well as conventional VL tasks such as Image Captioning and VQA. Experimental results showcase Shikra's promising performance. Furthermore, it enables numerous exciting applications, like providing mentioned objects' coordinates in chains of thoughts and comparing user-pointed regions similarities. Our code, model and dataset are accessed at https://github.com/shikras/shikra.
hub tools
citation-role summary
background 23
baseline 4
dataset 1
citation-polarity summary
claims ledger
- abstract In human conversations, individuals can indicate relevant regions within a scene while addressing others. In turn, the other person can then respond by referring to specific regions if necessary. This natural referential ability in dialogue remains absent in current Multimodal Large Language Models (MLLMs). To fill this gap, this paper proposes an MLLM called Shikra, which can handle spatial coordinate inputs and outputs in natural language. Its architecture consists of a vision encoder, an alignment layer, and a LLM. It is designed to be straightforward and simple, without the need for extra
authors
co-cited works
representative citing papers
MTLA is a training-free, post-hoc confidence score for multimodal LLM localization that restricts attention aggregation to the model's own predicted region and tokens, substantially improving hallucination detection and re-ranking across image, video, and audio.
DeepGaze3.5-VL treats visual scanpaths as discrete token sequences predicted autoregressively by vision-language models, achieving 2.18 bits IG on MIT1003 and outperforming prior specialized models even with matched encoders.
HG-Bench supplies 500 human-annotated homework samples and a page-aware protocol that measures complete-answer localization (FA) and step-level decomposition (FSm), exposing that no zero-shot VLM exceeds 55% on either metric.
Earth-OneVision is a unified 2B-parameter RS-MLLM supporting six modalities and nine tasks via FGVLA, SLIS, and PCMA mechanisms plus a 34M QA-pair dataset, reporting competitive or superior benchmark results versus larger models.
Video MLLMs show higher jailbreak rates with multi-clip videos than images or static videos, with success increasing alongside clip count and contextual diversity.
CAS mitigates object hallucinations in MLLMs by extracting two context preference vectors from designed conflict samples and applying signed residual injection at mid-early MLP layers without retraining or added latency.
OmniGF adapts VLMs via dual-branch decoding and head embeddings to unify precise multi-person gaze localization with semantic and social reasoning, claiming new SOTA on benchmarks.
IQA-Spider unifies reasoning, grounding, and referring for multi-granularity image quality assessment via a four-task paradigm and two-stage LMM training with training-free text-to-point mapping.
Qwen3-VL-Seg decodes MLLM bounding boxes into pixel-level referring segmentation via a lightweight box-guided mask decoder, new SA1B-ORS training data, and ORS-Bench evaluation, showing strong open-world performance.
RLFSeg repurposes pretrained generative models via Rectified Flow for direct latent-space image-to-mask mapping in text-based segmentation, outperforming diffusion-based methods especially in zero-shot cases.
CGC improves fine-grained multi-image understanding in MLLMs by constructing contrastive training instances from existing single-image annotations and adding a rule-based spatial reward, achieving SOTA on MIG-Bench and VLM2-Bench with transfer gains to other multimodal tasks.
By drawing object boxes and motion trails visually on video frames instead of serializing coordinates as text, BoxTuning reduces token costs dramatically and improves accuracy on video question answering benchmarks.
STORM is an end-to-end MLLM for referring multi-object tracking that uses task-composition learning to leverage sub-task data and introduces the STORM-Bench dataset, achieving SOTA results.
Suppressing low-attention tokens during the focus phase of vision-encoder processing reduces object hallucinations in LVLMs while preserving caption quality and adding negligible inference time.
Topo-R1 fine-tunes a vision-language model using a topology-aware reward and GRPO to detect anomalies such as broken or spurious connections in tubular segmentation masks, outperforming standard VLMs.
The paper delivers the first comprehensive review and unified taxonomy of agentic AI in remote sensing, covering single-agent copilots, multi-agent systems, planning mechanisms, benchmarks, and a roadmap while noting limitations in grounding and safety.
4D-RGPT uses perceptual 4D distillation to boost region-level 4D perception in multimodal LLMs and reports gains on existing and new video QA benchmarks.
PyramidDrop accelerates LVLMs by staged, similarity-based dropping of visual tokens that become redundant in deeper layers, delivering 40% faster training and 55% lower inference cost with comparable accuracy.
Q-Align trains LMMs on discrete text-defined levels for visual scoring, achieving SOTA on IQA, IAA, and VQA while unifying the tasks in OneAlign.
Large vision-language models exhibit severe object hallucination that varies with training instructions, and the proposed POPE polling method evaluates it more stably and flexibly than prior approaches.
GeoSearcher introduces anchor-centric reasoning supervised fine-tuning and process-faithful group relative policy optimization to improve MLLM-based remote sensing visual grounding.
SatAgent is a UAV-satellite collaborative spatial reasoning model using geometric 3D encoding, multi-view alignment, and a new 130K dataset that reports 25.91% and 11.69% gains over general and specialized baselines.
VisReflect generates continuous latent visual reflections to emphasize relevant visual features and guide attention in LVLMs, yielding 4.1% gains on image benchmarks and 1.8% on video benchmarks with 44% less inference time than zooming methods.
citing papers explorer
-
Vision as Unified Multimodal Generation
A single unified multimodal model matches leading task-specialized vision systems across detection, segmentation, dense geometry, and multi-view 3D by casting all outputs as native text or image generation.
-
Propose and Attend: Training-free MLLM Grounding Confidence via Multi-Token Localized Attention
MTLA is a training-free, post-hoc confidence score for multimodal LLM localization that restricts attention aggregation to the model's own predicted region and tokens, substantially improving hallucination detection and re-ranking across image, video, and audio.
-
DeepGaze3.5-VL: Modeling Scanpaths via Autoregressive Token Prediction
DeepGaze3.5-VL treats visual scanpaths as discrete token sequences predicted autoregressively by vision-language models, achieving 2.18 bits IG on MIT1003 and outperforming prior specialized models even with matched encoders.
-
HG-Bench: A Benchmark for Multi-Page Handwritten Answer-Region Grounding in Automated Homework Assessment
HG-Bench supplies 500 human-annotated homework samples and a page-aware protocol that measures complete-answer localization (FA) and step-level decomposition (FSm), exposing that no zero-shot VLM exceeds 55% on either metric.
-
Earth-OneVision: Extending Remote Sensing Multimodal Large Language Models to More Sensor Modalities and Tasks
Earth-OneVision is a unified 2B-parameter RS-MLLM supporting six modalities and nine tasks via FGVLA, SLIS, and PCMA mechanisms plus a 34M QA-pair dataset, reporting competitive or superior benchmark results versus larger models.
-
Jailbreaking Multimodal Large Language Models using Multi-Clip Video
Video MLLMs show higher jailbreak rates with multi-clip videos than images or static videos, with success increasing alongside clip count and contextual diversity.
-
Rethinking Visual Neglect: Steering via Context-Preference for MLLM Hallucination Mitigation
CAS mitigates object hallucinations in MLLMs by extracting two context preference vectors from designed conflict samples and applying signed residual injection at mid-early MLP layers without retraining or added latency.
-
OmniGF: A Dual-Branch Vision-Language Framework for Unified Gaze Following
OmniGF adapts VLMs via dual-branch decoding and head embeddings to unify precise multi-person gaze localization with semantic and social reasoning, claiming new SOTA on benchmarks.
-
IQA-Spider: Unifying Multi-Granularity Image Quality Assessment with Reasoning, Grounding and Referring
IQA-Spider unifies reasoning, grounding, and referring for multi-granularity image quality assessment via a four-task paradigm and two-stage LMM training with training-free text-to-point mapping.
-
Qwen3-VL-Seg: Unlocking Open-World Referring Segmentation with Vision-Language Grounding
Qwen3-VL-Seg decodes MLLM bounding boxes into pixel-level referring segmentation via a lightweight box-guided mask decoder, new SA1B-ORS training data, and ORS-Bench evaluation, showing strong open-world performance.
-
From Diffusion to Rectified Flow: Rethinking Text-Based Segmentation
RLFSeg repurposes pretrained generative models via Rectified Flow for direct latent-space image-to-mask mapping in text-based segmentation, outperforming diffusion-based methods especially in zero-shot cases.
-
CGC: Compositional Grounded Contrast for Fine-Grained Multi-Image Understanding
CGC improves fine-grained multi-image understanding in MLLMs by constructing contrastive training instances from existing single-image annotations and adding a rule-based spatial reward, achieving SOTA on MIG-Bench and VLM2-Bench with transfer gains to other multimodal tasks.
-
BoxTuning: Directly Injecting the Object Box for Multimodal Model Fine-Tuning
By drawing object boxes and motion trails visually on video frames instead of serializing coordinates as text, BoxTuning reduces token costs dramatically and improves accuracy on video question answering benchmarks.
-
STORM: End-to-End Referring Multi-Object Tracking in Videos
STORM is an end-to-end MLLM for referring multi-object tracking that uses task-composition learning to leverage sub-task data and introduces the STORM-Bench dataset, achieving SOTA results.
-
Focus Matters: Phase-Aware Suppression for Hallucination in Vision-Language Models
Suppressing low-attention tokens during the focus phase of vision-encoder processing reduces object hallucinations in LVLMs while preserving caption quality and adding negligible inference time.
-
Topo-R1: Detecting Topological Anomalies via Vision-Language Models
Topo-R1 fine-tunes a vision-language model using a topology-aware reward and GRPO to detect anomalies such as broken or spurious connections in tubular segmentation masks, outperforming standard VLMs.
-
Agentic AI in Remote Sensing: Foundations, Taxonomy, and Emerging Systems
The paper delivers the first comprehensive review and unified taxonomy of agentic AI in remote sensing, covering single-agent copilots, multi-agent systems, planning mechanisms, benchmarks, and a roadmap while noting limitations in grounding and safety.
-
4D-RGPT: Toward Region-level 4D Understanding via Perceptual Distillation
4D-RGPT uses perceptual 4D distillation to boost region-level 4D perception in multimodal LLMs and reports gains on existing and new video QA benchmarks.
-
PyramidDrop: Accelerating Your Large Vision-Language Models via Pyramid Visual Redundancy Reduction
PyramidDrop accelerates LVLMs by staged, similarity-based dropping of visual tokens that become redundant in deeper layers, delivering 40% faster training and 55% lower inference cost with comparable accuracy.
-
Q-Align: Teaching LMMs for Visual Scoring via Discrete Text-Defined Levels
Q-Align trains LMMs on discrete text-defined levels for visual scoring, achieving SOTA on IQA, IAA, and VQA while unifying the tasks in OneAlign.
-
Evaluating Object Hallucination in Large Vision-Language Models
Large vision-language models exhibit severe object hallucination that varies with training instructions, and the proposed POPE polling method evaluates it more stably and flexibly than prior approaches.
-
GeoSearcher: Anchor-Guided Progressive Reasoning for Remote Sensing Visual Grounding with Process Supervision
GeoSearcher introduces anchor-centric reasoning supervised fine-tuning and process-faithful group relative policy optimization to improve MLLM-based remote sensing visual grounding.
-
AeroVerse-SatAgent: UAV-Satellite Collaborative Spatial Reasoning Inspired by the Dual Visual Pathway Theory of Cognitive Neuroscience
SatAgent is a UAV-satellite collaborative spatial reasoning model using geometric 3D encoding, multi-view alignment, and a new 130K dataset that reports 25.91% and 11.69% gains over general and specialized baselines.
-
VisReflect: Latent Visual Reflection for Fine-Grained Perception in Long Visual Context
VisReflect generates continuous latent visual reflections to emphasize relevant visual features and guide attention in LVLMs, yielding 4.1% gains on image benchmarks and 1.8% on video benchmarks with 44% less inference time than zooming methods.
-
MotionAtlas: Detailed Region Captioning for Motion-Centric Videos
MotionAtlas supplies a 2,073-question benchmark, a self-bootstrap pipeline yielding 159k captions, and fine-tuned Video-MLLMs that deliver 5.2-point gains over Qwen3-VL-4B on motion tasks.
-
Enhancing Part-Level Point Grounding for Any Open-Source MLLMs
A plug-in Q-Synth Module plus Attention-to-Point Decoder converts text-conditioned attention in frozen MLLMs into point heatmaps, improving part-level grounding accuracy on multiple datasets.
-
HKVLM: Faithful Query--Region Binding for Frozen-Detector Visual Grounding
Under frozen perception, an explicit query-region alignment hook plus perception-grounded abstention reduces conditional binding errors and hallucinations, with an exact Acc=(1-SeeErr)(1-SayErr) decomposition.
-
Omni-Perception Policy Optimization for Multimodal Emotion Reasoning
A cue-coverage reward plus a modality-token KL penalty makes multimodal emotion-reasoning models cite more real visual/audio evidence and hallucinate less, with reported SoTA on emotion benchmarks.
-
Look Before You Zoom: Adaptive Routing for the Resolution-Context Trade-off in Visual RAG
ViRGo adaptively routes visual retrieval decisions in VLMs by estimating object scale from intrinsic localization heads combined with token confidence, matching patch retrieval on small objects, attention retrieval on large ones, and global processing when unnecessary.
-
HPP: Hierarchical Programmatic Probing for Long Video Understanding by Decoupling Perception and Reasoning
HPP decouples perception from reasoning in long-video VLMs by having an LLM run iterative programmatic probes on hierarchically segmented video, reporting gains on LongVideoBench, EgoSchema, VideoMME, and MLVU.
-
PerceptionDLM: Parallel Region Perception with Multimodal Diffusion Language Models
PerceptionDLM enables parallel region captioning in multimodal diffusion language models via prompting and attention masking, introduces ParaDLC-Bench, and claims first parallel region perception with DLMs.
-
Qwen-RobotWorld Technical Report: Unifying Embodied World Modeling through Language-Conditioned Video Generation
Qwen-RobotWorld is a language-conditioned video world model using Double-Stream MMDiT, an 8.6M-frame embodied corpus, and progressive curriculum training that ranks first on EWMBench and DreamGen Bench.
-
RoboPIN: Grounded Embodied Reasoning via Pinned Chain-of-Thought
Training a 4B vision-language model to emit identity-tracked, visually grounded reasoning anchors improves embodied spatial, multi-view, and pointing task performance over 7B baselines.
-
Iterative Visual Thinking and the Self-Correction Mirage in VLM Grounding
Iterative visual self-correction gains in VLM grounding are a mirage from oracle selection of the best trajectory step; under label-free stopping rules, iteration does not improve accuracy.
-
ShotCrop$^3$: Cropping Human-Centric Images into Cinematic Triple-Shot Compositions
ShotCrop uses three-stage training (CoT SFT, pseudo-label semi-supervised, GRPO-S) to produce triple-shot compositions and reports 2.82x better shot localization than GPT-5 on a 1.2k expert benchmark.
-
Self-Improving Small Object Grounding in LVLMs
Attention maps in LVLMs enable an IoU regressor (Pearson r > 0.67) and a training-free entropy-based selector that improves small-object localization by up to 19% on COCO and Objects365.
-
Through the PRISM: Principle-Aware, Interpretable, and Multi-Scale Evaluation of Visual Designs
PRISM benchmark perturbs Crello layouts into 110K samples isolating design principle violations, reveals limited sensitivity in several multimodal models, and proposes a multi-scale framework combining scorers, instruction-tuned VLMs, and prompt methods for interpretable design assessment.
-
Correcting Visual Blur Induced by Attention Distraction to Reduce Hallucinations: Algorithm and Theory
AFIP is a training-free attention-correction method that cuts object hallucination rates in multimodal LLMs by concentrating cross-head attention and restoring faded visual attention.
-
See What I Mean: Aligning Vision and Language Representations for Video Fine-grained Object Understanding
SWIM aligns cross-attention maps from object nouns to ground-truth masks during training on the new NL-Refer dataset to enable text-only fine-grained video object understanding in MLLMs.
-
A More Word-like Image Tokenization for MLLMs
DiVT clusters patch embeddings into coherent semantic units and adapts token count to image complexity, matching or exceeding baselines with fewer visual tokens on multimodal benchmarks.
-
WOW-Seg: A Word-free Open World Segmentation Model
WOW-Seg proposes a word-free open-world segmentation model using Mask2Token and Cascade Attention Mask modules, reporting 89.7 semantic similarity and 82.4 semantic IoU on LVIS with one-eighth the parameters of prior SOTA plus a new 7,662-class benchmark.
-
SceneParser: Hierarchical Scene Parsing for Visual Semantics Understanding
SceneParser introduces hierarchical scene parsing as object-part-affordance chains, a VLM trained with pseudo labels and curriculum learning, and SceneParser-Bench with 1.74M affordance annotations, showing better structure-aware results than existing MLLMs.
-
Guide, Think, Act: Interactive Embodied Reasoning in Vision-Language-Action Models
GTA-VLA lets humans steer a robot policy with spatial cues (points, boxes, traces) that condition the model's visual chain-of-thought, improving OOD robustness and recovering about 20% of failed episodes.
-
MMCL-Bench: Multimodal Context Learning from Visual Rules, Procedures, and Evidence
MMCL-Bench shows that even the strongest frontier multimodal models solve fewer than one-third of tasks requiring recovery and application of visual rules, procedures, and empirical patterns.
-
Mitigating Action-Relation Hallucinations in LVLMs via Relation-aware Visual Enhancement
A new attention-enhancement method using ARS scores and RVE reduces action-relation hallucinations in LVLMs while generalizing to spatial and object hallucinations.
-
Keep What Audio Cannot Say: Context-Preserving Token Pruning for Omni-LLMs
ContextGuard prunes 55% of tokens in Qwen2.5-Omni 7B while matching full performance on five of six audio-visual benchmarks by preserving audio-irrecoverable visual context.
-
SpatialForge: Bootstrapping 3D-Aware Spatial Reasoning from Open-World 2D Images
SpatialForge synthesizes 10 million spatial QA pairs from in-the-wild 2D images to train VLMs for better depth ordering, layout, and viewpoint-dependent reasoning.
-
Vocabulary Hijacking in LVLMs: Unveiling Critical Attention Heads by Excluding Inert Tokens to Mitigate Hallucination
LVLMs show vocabulary hijacking by inert tokens that decode to hijacking anchors; HABI locates them, NHAR finds resilient heads, and HAVAE boosts those heads to cut hallucinations.
-
VISTA: Video Interaction Spatio-Temporal Analysis Benchmark
VISTA is a new ~12K-pair benchmark and taxonomy for open-set multi-entity spatio-temporal understanding in VLMs that decomposes videos into entities, actions, and relational dynamics for multi-axis diagnostics.
-
State Beyond Appearance: Diagnosing and Improving State Consistency in Dial-Based Measurement Reading
MLLMs ignore dial state geometry and cluster by appearance, causing inconsistency under variations; TriSCA's state-distance alignment, metadata supervision, and objective alignment improve robustness on clock and gauge benchmarks.