Seeing from Another Perspective: Evaluating Multi-View Understanding in MLLMs
read the original abstract
Multi-view understanding, the ability to reconcile visual information across diverse viewpoints for effective navigation, manipulation, and 3D scene comprehension, is a fundamental challenge in Multi-Modal Large Language Models (MLLMs) to be used as embodied agents. While recent MLLMs have shown impressive advances in high-level reasoning and planning, they frequently fall short when confronted with multi-view geometric consistency and cross-view correspondence. To comprehensively evaluate the challenges of MLLMs in multi-view scene reasoning, we propose All-Angles Bench, a benchmark of over 2,100 human carefully annotated multi-view question-answer pairs across 90 diverse real-world scenes. Our six tasks (counting, attribute identification, relative distance, relative direction, object manipulation, and camera pose estimation) specifically test model's geometric correspondence and the capacity to align information consistently across views. Our extensive experiments, benchmark on 27 representative MLLMs including Gemini-2.0-Flash, Claude-3.7-Sonnet, and GPT-4o against human evaluators reveals a substantial performance gap, indicating that current MLLMs remain far from human-level proficiency. Through in-depth analysis, we show that MLLMs are particularly underperforming under two aspects: (1) cross-view correspondence for partially occluded views and (2) establishing the coarse camera poses. These findings highlight the necessity of domain-specific refinements or modules that embed stronger multi-view awareness. We believe that our All-Angles Bench offers valuable insights and contribute to bridging the gap between MLLMs and human-level multi-view understanding. The project and benchmark are publicly available at https://danielchyeh.github.io/All-Angles-Bench/.
This paper has not been read by Pith yet.
Forward citations
Cited by 28 Pith papers
-
TriViewBench: Controlled Complexity Scaling for Multi-View Structural Reasoning in MLLMs
TriViewBench shows all tested MLLMs follow the same capability order with sharp drops on complex multi-view tasks and near-zero gain from Chain-of-Thought prompting.
-
SSMNBench: Diagnosing Image-based Cross-View Human-Object Understanding via Single-View Sufficiency and Multi-View Necessity
SSMNBench shows that MLLMs suffer distraction degradation on single-view-sufficient tasks and fail to integrate geometric evidence across views, instead relying on semantic averaging and view preference.
-
SATURN: Symbolic Spatial Reasoning for Multi-Perspective Grounding
SATURN reconstructs approximate 3D scenes, derives soft perspective-aware predicates, and executes them symbolically to achieve stable performance on complex multi-perspective spatial grounding tasks where VLMs degrade.
-
Eliciting Complex Spatial Reasoning in MLLMs through Wide-Baseline Matching
Authors create ReasonMatch-Bench and DCRL training to boost MLLM performance on wide-baseline matching, reporting gains over baselines while preserving general capabilities.
-
Consistent Yet Wrong: Evidence Insensitivity in Spatial Vision-Language Models
VLMs produce highly consistent but inaccurate answers to metric spatial queries across views, indicating prior-driven collapse rather than evidence-sensitive reasoning.
-
X-Stream: Exploring MLLMs as Multiplexers for Multi-Stream Understanding
X-Stream benchmark shows state-of-the-art MLLMs achieve only about 50% on multi-stream video tasks and exhibit poor proactive ability.
-
X-Stream: Exploring MLLMs as Multiplexers for Multi-Stream Understanding
X-Stream benchmark shows SOTA MLLMs score ~50% on concurrent multi-stream tasks and lack proactive ability, using a dual-verification pipeline to avoid single-stream bias.
-
DRIVESPATIAL: A Benchmark for Spatiotemporal Intelligence in VLMs for Autonomous Driving
DriveSpatial benchmark shows the best of 15 VLMs trails humans by 28.4 points on spatiotemporal driving tasks, with cognitive scene construction as the main failure mode.
-
DRIVESPATIAL: A Benchmark for Spatiotemporal Intelligence in VLMs for Autonomous Driving
DriveSpatial benchmark shows the strongest of 15 VLMs trails humans by 28.4 points on spatiotemporal tasks, with cognitive scene construction as the primary weakness.
-
SCP: Spatial Causal Prediction in Video
SCP defines a new benchmark task for predicting spatial causal outcomes beyond direct observation and shows that 23 leading models lag far behind humans on it.
-
SpatialMosaic: A Multiview VLM Dataset for Partial Visibility
SpatialMosaic introduces a 2M-pair multi-view QA dataset and 1M-pair benchmark for MLLMs on spatial reasoning under partial visibility, plus a hybrid baseline that integrates 3D reconstruction models as geometry encoders.
-
Seeing Isn't Orienting: A Cognitively Grounded Benchmark Reveals Systematic Orientation Failures in MLLMs
DORI benchmark shows top vision-language models reach only 54.2% accuracy on coarse orientation tasks and 33% on granular judgments, with sharp drops on reference-frame shifts and compound rotations.
-
Vesta: A Generalist Embodied Reasoning Model
Vesta is a unified embodied generalist model that outperforms specialist baselines by over 20% on average and improves real-world robotic task success by over 35%.
-
Imaginative Perception Tokens Enhance Spatial Reasoning in Multimodal Language Models
IPT supervision improves spatial reasoning in VLMs on perspective taking, path tracing, and multiview counting tasks, often outperforming textual chain-of-thought while remaining consistent with observed inputs.
-
Consistent Yet Wrong: Evidence Insensitivity in Spatial Vision-Language Models
Leading VLMs show high cross-view consistency paired with low metric accuracy on distance queries, indicating evidence-insensitive reasoning rather than geometric grounding.
-
Beyond 3D VQAs: Injecting 3D Spatial Priors into Vision-Language Models for Enhanced Geometric Reasoning
GASP injects geometric priors into VLMs via a deep-supervised correspondence head trained on video point correspondences and depth consistency, raising internal matching accuracy and delivering gains on spatial benchm...
-
Cambrian-P: Pose-Grounded Video Understanding
Cambrian-P adds per-frame camera pose tokens and a regression head to video MLLMs, delivering 4.5-6.5% gains on spatial benchmarks, generalization to other video QA tasks, and SOTA streaming pose estimation on ScanNet.
-
Cambrian-S: Towards Spatial Supersensing in Video
Cambrian-S introduces VSI-SUPER benchmarks for long-horizon spatial recall and counting, shows data scaling yields 30% gains on existing tests, and demonstrates a self-supervised next-latent predictor using surprise o...
-
GLM-4.5V and GLM-4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning
GLM-4.5V reaches state-of-the-art results on 42 multimodal benchmarks among open-source models of similar size by applying reinforcement learning with curriculum sampling to a strong vision foundation model.
-
LongSpace: Exploring Long-Horizon Spatial Memory from Perception to Recall in Video
Presents LongSpace-Bench benchmark and LongSpace framework that chunks long videos, adds 3D structural cues, and builds layer-aware memory to improve spatial reasoning in multimodal LLMs.
-
CrossView Suite: Harnessing Cross-view Spatial Intelligence of MLLMs with Dataset, Model and Benchmark
CrossView Suite supplies a 1.6M-sample dataset, scene-disjoint benchmark, and explicit-alignment framework to advance MLLMs from single-view perception to cross-view spatial intelligence.
-
From Where Things Are to What They Are For: Benchmarking Spatial-Functional Intelligence in Multimodal LLMs
SFI-Bench shows current multimodal LLMs struggle to integrate spatial memory with functional reasoning and external knowledge in video tasks.
-
OpenSpatial: A Principled Data Engine for Empowering Spatial Intelligence
OpenSpatial supplies a principled open-source data engine and 3-million-sample dataset that raises spatial-reasoning model performance by an average of 19 percent on benchmarks.
-
Kwai Keye-VL-2.0 Technical Report
Kwai Keye-VL-2.0-30B-A3B is a 30B MoE model with 3B active parameters using DSA adaptation and MOPD distillation that reports SOTA results on video understanding and agent benchmarks.
-
JoyAI-Image: Awaking Spatial Intelligence in Unified Multimodal Understanding and Generation
JoyAI-Image unifies visual understanding and generation via an MLLM-MMDiT architecture with spatial training signals to reach competitive benchmark performance and stronger spatial intelligence.
-
JoyAI-Image: Awaking Spatial Intelligence in Unified Multimodal Understanding and Generation
JoyAI-Image unifies visual understanding, generation, and editing in one model and claims stronger spatial intelligence through bidirectional perception-generation loops.
-
Seed1.5-VL Technical Report
Seed1.5-VL is a compact multimodal model that sets new records on dozens of vision-language benchmarks and outperforms prior systems on agent-style tasks.
-
Seed2.0 Model Card: Towards Intelligence Frontier for Real-World Complexity
Seed2.0 model series reports gains in reasoning, visual understanding, search, and reliability on intricate long-horizon tasks via an internal evaluation system.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.