Research map / Video / Video-language understanding
Looking Beyond the Obvious: A Survey on Abstract Concept Recognition for Video Understanding
Gowreesh Mago, Pascal Mettes, Stevan Rudinac
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 |
|---|---|---|---|
| map abstract fashion words to concrete concepts via PCA in the latent space | VLM zero-shot retrieval · Fashion abstract-word retrieval | retrieval accuracy (%): 31.1% vs. 43.7% | Improved |
| add multimodal (visual) input instead of text only (ablation) | MIntRec 2.0 intent model · MIntRec 2.0 intent classification | F1: increase in F1-score by 3.57 points | Improved |
| use multimodal input for video ads (ablation) | LLaVA-Video · AdsQA | accuracy (%): from 66.58% to 71.73% | Improved |
| use supervised multimodal model instead of zero-shot GPT-4 on transcripts (ablation) | ad topic categorization · Bose et al. ads benchmark topic task | accuracy: falls behind by 32% | Improved |
| fuse text, audio and vision rather than single modality (ablation) | ad understanding baselines · Bose et al. ads benchmark | task performance | Improved |
| add longer chain-of-thought reasoning (ablation) | VLMs on AdsQA · AdsQA | accuracy | Hurt |
| add hypothesis generation and supervised signals to an LLM | Claude Sonnet 3.5 · GPVD | accuracy (%): 85.5% accuracy, outperforming supervised baselines at 80% | Improved |
| add chain-of-thought prompting (ablation) | Foundation models on Oops!-based tasks · Oops!-based unintentional action tasks | accuracy | No clear effect |
Closest papers by meaning
- Look, Remember and Reason: Grounded Reasoning in Videos with Language Models (ICLR 2024 · similarity 0.87)
- PerceptionLM: Open-Access Data and Models for Detailed Visual Understanding (NeurIPS 2025 · similarity 0.87)
- OmniViD: A Generative Framework for Universal Video Understanding (CVPR 2024 · similarity 0.87)
- MMBench-Video: A Long-Form Multi-Shot Benchmark for Holistic Video Understanding (NeurIPS 2024 · similarity 0.87)
- ConViS-Bench: Estimating Video Similarity Through Semantic Concepts (NeurIPS 2025 · similarity 0.87)
- HENASY: Learning to Assemble Scene-Entities for Interpretable Egocentric Video-Language Model (NeurIPS 2024 · similarity 0.87)
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.
