Subtopic Deep Dive
Affective Computing Databases
Research Guide
What is Affective Computing Databases?
Affective Computing Databases are standardized, annotated datasets capturing emotional expressions across physiological, audio-visual, and facial modalities to enable reproducible emotion recognition research.
Key datasets include DEAP for EEG-based emotions (Zheng and Lu, 2015), RAVDESS for audio-visual speech and song (Livingstone and Russo, 2018), and IEMOCAP for interactive emotional speech. These databases feature elicitation protocols like music listening (Kim and André, 2008) and crowd-sourced annotations (Cao et al., 2014). Over 20 major databases exist, with foundational works pre-2015 cited over 700 times each.
Why It Matters
Standardized databases like WESAD (Schmidt et al., 2018) support wearable stress detection in health monitoring, enabling models with 90%+ accuracy via multimodal signals. RAF-DB (Li et al., 2017) benchmarks real-world facial expression recognition for surveillance applications. CREMA-D (Cao et al., 2014) aids multimodal actor analysis in human-computer interaction, with RAVDESS (Livingstone and Russo, 2018) used in 1,600+ studies for speech emotion AI in virtual assistants.
Key Research Challenges
Annotation Reliability
Crowd-sourced labels in RAF-DB vary due to subjective emotion perception (Li et al., 2017). Spontaneous micro-expressions in CASME II require expert validation for subtlety (Yan et al., 2014). Achieving inter-annotator agreement above 0.8 remains difficult across modalities.
Multimodal Synchronization
Aligning EEG, audio, and video in WESAD demands precise timestamps during stress elicitation (Schmidt et al., 2018). Physiological signals in music-listening protocols drift out of sync (Kim and André, 2008). This hampers deep learning models like DGCNN (Song et al., 2018).
Dataset Generalization
Models trained on DEAP EEG fail on diverse populations due to frequency band variations (Zheng and Lu, 2015). Wild expressions in RAF-DB expose lab-dataset biases (Li et al., 2017). Scaling to real-world via dynamical graphs needs stable patterns (Zheng et al., 2017).
Essential Papers
Investigating Critical Frequency Bands and Channels for EEG-Based Emotion Recognition with Deep Neural Networks
Wei‐Long Zheng, Bao‐Liang Lu · 2015 · IEEE Transactions on Autonomous Mental Development · 2.2K citations
To investigate critical frequency bands and channels, this paper introduces deep belief networks (DBNs) to constructing EEG-based emotion recognition models for three emotions: positive, neutral an...
The Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS): A dynamic, multimodal set of facial and vocal expressions in North American English
Steven R. Livingstone, Frank Russo · 2018 · PLoS ONE · 1.7K citations
The RAVDESS is a validated multimodal database of emotional speech and song. The database is gender balanced consisting of 24 professional actors, vocalizing lexically-matched statements in a neutr...
Enhancements in Immediate Speech Emotion Detection: Harnessing Prosodic and Spectral Characteristics
Z. B. M. D. Shah, SHAN Zhiyong, Adnan · 2024 · International Journal of Innovative Science and Research Technology (IJISRT) · 1.7K citations
Speech is essential to human communication for expressing and understanding feelings. Emotional speech processing has challenges with expert data sampling, dataset organization, and computational c...
EEG Emotion Recognition Using Dynamical Graph Convolutional Neural Networks
Tengfei Song, Wenming Zheng, Peng Song et al. · 2018 · IEEE Transactions on Affective Computing · 1.4K citations
In this paper, a multichannel EEG emotion recognition method based on a novel dynamical graph convolutional neural networks (DGCNN) is proposed. The basic idea of the proposed EEG emotion recogniti...
Reliable Crowdsourcing and Deep Locality-Preserving Learning for Expression Recognition in the Wild
Shan Li, Weihong Deng, Junping Du · 2017 · 1.2K citations
Past research on facial expressions have used relatively limited datasets, which makes it unclear whether current methods can be employed in real world. In this paper, we present a novel database, ...
Emotion recognition based on physiological changes in music listening
Jonghwa Kim, Elisabeth André · 2008 · IEEE Transactions on Pattern Analysis and Machine Intelligence · 1.1K citations
Little attention has been paid so far to physiological signals for emotion recognition compared to audiovisual emotion channels such as facial expression or speech. This paper investigates the pote...
Introducing WESAD, a Multimodal Dataset for Wearable Stress and Affect Detection
Philip Schmidt, Attila Reiss, Robert Duerichen et al. · 2018 · 1.0K citations
Affect recognition aims to detect a person's affective state based on observables, with the goal to e.g. improve human-computer interaction. Long-term stress is known to have severe implications on...
Reading Guide
Foundational Papers
Start with Kim and André (2008) for physiological signal potential in music tasks; Cao et al. (2014) CREMA-D for multimodal crowd-sourcing; Yan et al. (2014) CASME II for micro-expression annotation standards.
Recent Advances
Study Zheng and Lu (2015) DEAP analysis (2206 cites); Livingstone and Russo (2018) RAVDESS (1662 cites); Schmidt et al. (2018) WESAD for wearables.
Core Methods
Elicitation protocols (music, acting), crowd/expert annotation, synchronization of EEG/audio/video, benchmarking with DBNs (Zheng and Lu, 2015) or DGCNN (Song et al., 2018).
How PapersFlow Helps You Research Affective Computing Databases
Discover & Search
Research Agent uses searchPapers and citationGraph to map 50+ papers citing RAVDESS (Livingstone and Russo, 2018), revealing DEAP extensions. exaSearch queries 'EEG emotion databases post-2015' for WESAD variants, while findSimilarPapers links CASME II to RAF-DB benchmarks.
Analyze & Verify
Analysis Agent runs readPaperContent on WESAD (Schmidt et al., 2018) to extract annotation protocols, then verifyResponse with CoVe checks claims against raw metadata. runPythonAnalysis loads DEAP EEG stats via pandas for frequency band verification (Zheng and Lu, 2015), graded by GRADE for evidence strength.
Synthesize & Write
Synthesis Agent detects gaps in micro-expression coverage between CASME II and RAF-DB, flagging contradictions in elicitation methods. Writing Agent applies latexEditText to benchmark tables, latexSyncCitations for 20+ refs, and latexCompile for publication-ready reviews; exportMermaid diagrams modality overlaps.
Use Cases
"Analyze DEAP EEG frequency bands for valence-arousal model"
Research Agent → searchPapers('DEAP dataset') → Analysis Agent → runPythonAnalysis(pandas on EEG channels, matplotlib power spectra) → statistical verification of Zheng and Lu (2015) bands → GRADE report with p-values.
"Benchmark RAVDESS vs CREMA-D for speech emotion LaTeX table"
Research Agent → citationGraph(RAVDESS) → Synthesis Agent → gap detection → Writing Agent → latexEditText(table), latexSyncCitations(20 refs), latexCompile → PDF with actor comparison metrics.
"Find GitHub repos implementing RAF-DB evaluation code"
Research Agent → paperExtractUrls(RAF-DB paper) → Code Discovery → paperFindGithubRepo → githubRepoInspect → runPythonAnalysis on sample scripts → verified wild-expression benchmarks.
Automated Workflows
Deep Research workflow scans 50+ papers on physiological datasets, chaining searchPapers → citationGraph → structured report ranking RAVDESS by citations. DeepScan applies 7-step analysis to CASME II, with CoVe checkpoints verifying micro-expression baselines. Theorizer generates hypotheses on WESAD multimodal fusion from literature patterns.
Frequently Asked Questions
What defines an Affective Computing Database?
Annotated datasets of emotional data from EEG, speech, facial, or physiological signals, designed for machine learning benchmarks like DEAP and RAVDESS.
What are common methods for dataset creation?
Elicitation via music (Kim and André, 2008), acting (Livingstone and Russo, 2018), or crowd-sourcing (Li et al., 2017), with expert validation for micro-expressions (Yan et al., 2014).
What are key papers?
Zheng and Lu (2015, 2206 cites) on DEAP EEG; Livingstone and Russo (2018, 1662 cites) on RAVDESS; Schmidt et al. (2018, 1039 cites) on WESAD.
What open problems exist?
Generalization across cultures, real-time synchronization of modalities, and scalable annotation for spontaneous wild expressions beyond RAF-DB.
Research Emotion and Mood Recognition with AI
PapersFlow provides specialized AI tools for Psychology researchers. Here are the most relevant for this topic:
Systematic Review
AI-powered evidence synthesis with documented search strategies
AI Literature Review
Automate paper discovery and synthesis across 474M+ papers
Find Disagreement
Discover conflicting findings and counter-evidence
Deep Research Reports
Multi-source evidence synthesis with counter-evidence
See how researchers in Social Sciences use PapersFlow
Field-specific workflows, example queries, and use cases.
Start Researching Affective Computing Databases with AI
Search 474M+ papers, run AI-powered literature reviews, and write with integrated citations — all in one workspace.
See how PapersFlow works for Psychology researchers
Part of the Emotion and Mood Recognition Research Guide