Research map / Vision-language and multimodal / Human and embodied modeling
Business-Aware Visual Concept Discovery from Social Media for Multimodal Business Venue Recognition
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| select training images by two-fold cross-validation instead of using all crawled images (ablation) | concept classifiers · Instagram concept classification | rank-1 accuracy: 36.5% -> 48.5% | Helped |
| select training images by cross-validation instead of random selection (ablation) | concept classifiers · Instagram concept classification | rank-1 accuracy: 38.7% -> 48.5% | Helped |
| add word-vector representation of top-5 detected concepts to BA-CNN (C+V) concept features (ablation) | BA-CNN (C+V) · venue recognition, 5 candidates | rank-1 accuracy | Helped |
| train word vectors on Yelp reviews instead of Google News (ablation) | WordVec text-only recognition · venue recognition, 2 candidates | rank-1 accuracy: 65.8% -> 69.1% | Helped |
| train word vectors on Yelp reviews instead of Google News | WordVec text-only recognition · venue recognition, 5 candidates | rank-1 accuracy: 39.1% -> 42.3% | Helped |
| add word-vector text modality to image concepts, fusing early | BA-CNN (C+V) · venue recognition, 2 candidates | rank-1 accuracy: 77.5% -> 78.5% | Helped |
| add database reviews when few database images per venue | BA-CNN (C+V) · venue recognition, sparse images (Figure 8) | rank-1 accuracy | Helped |
| use BA-CNN (C+V) hybrid concept+venue labels instead of concept-only or venue-only labels | CNN features · venue recognition | rank-1 accuracy | Helped |
Closest papers by meaning
- GeoReward: Mitigating Contextual Variable Overestimation in Vision-Language Models for Cross-Market Preference Prediction (ICML 2026 · similarity 0.82)
- Towards Implicit Aggregation: Robust Image Representation for Place Recognition in the Transformer Era (NeurIPS 2025 · similarity 0.82)
- Focus on Local: Finding Reliable Discriminative Regions for Visual Place Recognition (AAAI 2025 · similarity 0.81)
- BoQ: A Place is Worth a Bag of Learnable Queries (CVPR 2024 · similarity 0.81)
- Beyond Graph Convolution: Multimodal Recommendation with Topology-aware MLPs (AAAI 2025 · similarity 0.81)
- TIGER-FG: Text-Guided Implicit Fine-Grained Grounding for E-commerce Retrieval (NeurIPS 2026 · similarity 0.81)
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.