Research map / Vision-language and multimodal / Multimodal LLMs
Visual Chronicles: Using Multimodal LLMs to Analyze Massive Collections of Images
Boyang Deng, Songyou Peng, Kyle Genova, Gordon Wetzstein, Noah Snavely, Leonidas Guibas, Thomas Funkhouser
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use Gemini MLLM instead of embedding or feature distances | change detection · Change detection (NYC test) | AP: 16.44 % -> 76.56 % | Improved |
| use Gemini MLLM instead of NV-Emb for trend verification | trend verification · Trend verification (NYC test) | AP: 73.13 % -> 86.63 % | Improved |
| use hybrid embedding-then-MLLM verification instead of random MLLM verification | trend classification · Trend classification Acc.@50 | accuracy: 31.8% -> 93.9% | Improved |
| use hybrid verification instead of NV-Emb alone | trend classification · Trend classification Acc.@100 | accuracy: 69.6% -> 94.6% | Improved |
| use hybrid verification instead of NV-Emb alone | trend classification · Trend classification Acc.@200 | accuracy: 81.8% -> 98.3% | Improved |
| use Gemini instead of LLaVA (ablation) | change detection · Change detection (NYC test) | AP: 64.31 % -> 76.56 % | Improved |
| use Gemini instead of LLaVA (ablation) | trend verification · Trend verification (NYC test) | AP: 81.11 % -> 86.63 % | Improved |
| add Self-Critic verification round (ablation) | local change detection · Change detection (NYC test) | precision: 57.83 % -> 81.74 % | Improved |
| add Self-Critic verification round (ablation) | local change detection · Change detection (NYC test) | recall: 91.50 % -> 89.87 % | No clear effect |
| increase k from 2N to 5N (ablation) | hybrid trend verification · Trend classification Acc.@50 | accuracy: 88.6% -> 97.2% | Improved |
| increase k from 2N to 5N (ablation) | hybrid trend verification · Trend classification Acc.@100 | accuracy: 91.4% -> 97.1% | Improved |
| increase k from 2N to 5N (ablation) | hybrid trend verification · Trend classification Acc.@200 | accuracy: 95.7% -> 99.5% | Improved |
Closest papers by meaning
- DynamicVL: Benchmarking Multimodal Large Language Models for Dynamic City Understanding (NeurIPS 2025 · similarity 0.88)
- Seeking and Updating with Live Visual Knowledge (NeurIPS 2025 · similarity 0.85)
- The STVchrono Dataset: Towards Continuous Change Recognition in Time (CVPR 2024 · similarity 0.85)
- UrbanFeel:A Comprehensive Benchmark for Temporal and Perceptual Understanding of City Scenes through Human Perspective (ICLR 2026 · similarity 0.85)
- Video-MME: The First-Ever Comprehensive Evaluation Benchmark of Multi-modal LLMs in Video Analysis (CVPR 2025 · similarity 0.85)
- V?: Guided Visual Search as a Core Mechanism in Multimodal LLMs (CVPR 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
