An interactive web-based visualization platform for exploring multimodal emotion knowledge graphs, connecting sentiment words, facial expressions, social media emojis, and animal expressions.
Interactive knowledge graph visualization with dark mode
Interactive knowledge graph visualization with light mode
- 🖱️ Drag & Drop: Click and drag any node to reposition it - watch the graph dynamically adjust its layout
- 🔍 Zoom Control: Scroll to zoom in and out of the graph for detailed exploration
- 👆 Hover to Explore: Move your mouse over nodes to see which emotion category they belong to
- 🔗 Click to Connect: Click on any node to highlight its direct connections and relationships
- 🎭 Expression Images: View emotion representations across three modalities:
- 😊 Human Face Expressions - Real facial emotion images
- 😃 Social Media Emojis - Digital emotion representations
- 🐱 Animal Expressions - Emotion expressions from cats, dogs, and hamsters
- 🌓 Dark/Light Mode: Toggle between dark and light themes for comfortable viewing
The knowledge graph integrates multimodal emotional data representing 7 core emotions:
| Emotion | English | Description |
|---|---|---|
| 乐 | Joy | Happiness, pleasure, delight |
| 好 | Good | Positive feelings, satisfaction |
| 怒 | Anger | Frustration, rage, irritation |
| 哀 | Sadness | Sorrow, grief, melancholy |
| 惧 | Fear | Anxiety, worry, terror |
| 恶 | Disgust | Aversion, revulsion, dislike |
| 惊 | Surprise | Astonishment, amazement, shock |
Dataset Sources:
- Text Emotion Words: 27,466 Chinese emotion words from Dalian University of Technology
- Facial Expressions: JAFFE dataset with 216 images across 7 emotions
- Animal Expressions: 84 images (cats, dogs, hamsters) collected and curated
- Social Media Emojis: 77 emoji images from emojipedia.org
Bachelor's Thesis
- Topic: "Construction of a Multimodal Knowledge Graph for Fine-Grained Sentiment Analysis" (view the project website)
- Institution: Northeastern University (NEU), China
China Invention Patent (Granted)
- Title: Method for Constructing and Presenting a Multimodal Sentiment Knowledge Graph
- Patent Number: CN ZL202011319237.8
- Grant Date: September 7, 2021
- Status: Co-inventor
- Frontend: HTML5, CSS3, JavaScript
- Visualization: D3.js force-directed graph
- UI Framework: Bootstrap
- Data Storage: Neo4j graph database
- Ontology Editor: Protégé
If you use this work in your research, please cite:
@mastersthesis{li2021emotionnet,
title={Construction of a Multimodal Knowledge Graph for Fine-Grained Sentiment Analysis},
author={Li, Jiufeng},
year={2020},
school={Northeastern University},
address={Shenyang, China},
type={Bachelor's Thesis}
}Patent Citation:
Li, J. et al. (2021). Method for Constructing and Presenting a Multimodal Sentiment Knowledge Graph.
China Patent No. CN ZL202011319237.8.
This project represents academic research on multimodal sentiment analysis and knowledge graph construction.