💗 WIA Emotion AI Standard Ebook | Chapter 4 of 8

💗 Chapter 4: Phase 1 - Emotion Data Format

Hongik Ingan (弘益人間)

"Benefit All Humanity"

A standardized data format is the foundation of interoperability. The WIA Emotion AI data format ensures that emotion data can be shared and processed across any compliant system.

4.1 Overview

4.1.1 Purpose

Phase 1 defines the JSON-based data format for representing emotion analysis results. This standardized format enables:

Interoperability between different vendors
Consistent data storage and retrieval
Easy integration with existing systems
Clear semantics for all emotion data fields

4.1.2 Design Goals

Goal	Implementation
Simplicity	JSON format, clear field names
Completeness	Supports all emotion models and modalities
Extensibility	Custom fields allowed with prefixes
Validation	JSON Schema for automated verification
Precision	High-resolution timestamps and values

4.2 Base Message Structure

4.2.1 Complete Schema

{
    "$schema": "https://wiastandards.com/emotion-ai/v1/schema.json",
    "format": "WIA-EMOTION-AI-v1.0",
    "timestamp": "2025-12-19T10:30:00.000Z",
    "subject_id": "user_12345",
    "session_id": "sess_abc123",
    "modality": "facial",

    "emotions": {
        "primary": {
            "label": "happiness",
            "confidence": 0.87
        },
        "secondary": {
            "label": "surprise",
            "confidence": 0.23
        },
        "all": [
            { "label": "happiness", "confidence": 0.87 },
            { "label": "surprise", "confidence": 0.23 },
            { "label": "neutral", "confidence": 0.15 },
            { "label": "sadness", "confidence": 0.05 },
            { "label": "anger", "confidence": 0.03 },
            { "label": "fear", "confidence": 0.02 },
            { "label": "disgust", "confidence": 0.01 }
        ]
    },

    "dimensions": {
        "valence": 0.72,
        "arousal": 0.45,
        "dominance": 0.55
    },

    "action_units": [
        { "au": "AU6", "intensity": 0.8, "name": "Cheek Raiser" },
        { "au": "AU12", "intensity": 0.9, "name": "Lip Corner Puller" },
        { "au": "AU1", "intensity": 0.3, "name": "Inner Brow Raiser" }
    ],

    "face": {
        "detected": true,
        "bbox": { "x": 120, "y": 80, "width": 200, "height": 250 },
        "landmarks": 68,
        "pose": { "pitch": 5.2, "yaw": -3.1, "roll": 1.2 }
    },

    "metadata": {
        "model_version": "2.1.0",
        "model_name": "WIA-EmotionNet",
        "processing_time_ms": 45,
        "device_id": "cam_001",
        "culture_context": "global"
    }
}

4.3 Field Specifications

4.3.1 Root Fields

Field	Type	Required	Description
format	string	REQUIRED	Format identifier: "WIA-EMOTION-AI-v1.0"
timestamp	string (ISO 8601)	REQUIRED	Analysis timestamp
subject_id	string	OPTIONAL	Identifier for the subject (anonymized)
session_id	string	OPTIONAL	Session identifier for grouping
modality	string	REQUIRED	Input type: facial, voice, text, biosignal, multimodal

4.3.2 Emotions Object

Field	Type	Required	Description
emotions.primary	object	REQUIRED	Highest confidence emotion
emotions.primary.label	string	REQUIRED	Emotion label (happiness, sadness, etc.)
emotions.primary.confidence	number (0-1)	REQUIRED	Confidence score
emotions.secondary	object	OPTIONAL	Second highest confidence emotion
emotions.all	array	OPTIONAL	All emotion scores

4.3.3 Valid Emotion Labels

Basic Emotions (required support):
  "happiness"  - Joy, pleasure, contentment
  "sadness"    - Sorrow, grief, melancholy
  "anger"      - Displeasure, frustration, rage
  "fear"       - Apprehension, anxiety, terror
  "disgust"    - Revulsion, aversion
  "surprise"   - Astonishment, amazement
  "neutral"    - No strong emotion detected

Extended Emotions (optional):
  "contempt"   - Scorn, disdain
  "confusion"  - Uncertainty, puzzlement
  "interest"   - Curiosity, engagement
  "boredom"    - Disengagement
  "excitement" - High positive arousal
  "anxiety"    - Worry, unease

4.3.4 Dimensions Object

Field	Type	Range	Description
dimensions.valence	number	-1.0 to +1.0	Positive/negative feeling
dimensions.arousal	number	-1.0 to +1.0	Energy/activation level
dimensions.dominance	number	-1.0 to +1.0	Control/influence (optional)

4.3.5 Action Units Array

Field	Type	Description
action_units[].au	string	AU code (AU1, AU2, AU4, etc.)
action_units[].intensity	number (0-1)	Activation intensity
action_units[].name	string	Human-readable name (optional)

4.4 Modality-Specific Extensions

4.4.1 Facial Modality

{
    "modality": "facial",
    "face": {
        "detected": true,
        "count": 1,
        "bbox": {
            "x": 120,
            "y": 80,
            "width": 200,
            "height": 250
        },
        "landmarks": 68,
        "pose": {
            "pitch": 5.2,
            "yaw": -3.1,
            "roll": 1.2
        },
        "quality": {
            "blur": 0.1,
            "exposure": 0.8,
            "occlusion": 0.05
        }
    },
    "action_units": [
        { "au": "AU6", "intensity": 0.8 },
        { "au": "AU12", "intensity": 0.9 }
    ]
}

4.4.2 Voice Modality

{
    "modality": "voice",
    "voice": {
        "duration_ms": 3500,
        "sample_rate": 44100,
        "features": {
            "pitch_mean": 180.5,
            "pitch_std": 25.3,
            "intensity_mean": 65.2,
            "speech_rate": 4.2,
            "pause_ratio": 0.15
        },
        "language": "en-US",
        "transcript": "I'm really happy about this!"
    }
}

4.4.3 Text Modality

{
    "modality": "text",
    "text": {
        "content": "This product is absolutely amazing!",
        "language": "en",
        "length": 37,
        "sentiment": {
            "polarity": 0.92,
            "subjectivity": 0.85
        },
        "entities": [
            {
                "text": "product",
                "emotion": "happiness",
                "confidence": 0.88
            }
        ],
        "aspects": [
            {
                "aspect": "quality",
                "sentiment": 0.95
            }
        ]
    }
}

4.4.4 Biosignal Modality

{
    "modality": "biosignal",
    "biosignal": {
        "signals": ["ecg", "eda"],
        "duration_ms": 60000,
        "metrics": {
            "heart_rate": {
                "mean": 75,
                "std": 8,
                "hrv_rmssd": 42.5
            },
            "eda": {
                "scl_mean": 3.2,
                "scr_count": 5,
                "scr_amplitude_mean": 0.8
            }
        },
        "derived": {
            "stress_level": 0.35,
            "engagement": 0.72,
            "relaxation": 0.55
        }
    }
}

4.4.5 Multimodal

{
    "modality": "multimodal",
    "modalities_used": ["facial", "voice"],
    "fusion_method": "weighted_average",
    "modality_weights": {
        "facial": 0.6,
        "voice": 0.4
    },
    "modality_results": {
        "facial": {
            "emotions": { "primary": { "label": "happiness", "confidence": 0.85 } },
            "dimensions": { "valence": 0.7, "arousal": 0.5 }
        },
        "voice": {
            "emotions": { "primary": { "label": "happiness", "confidence": 0.78 } },
            "dimensions": { "valence": 0.6, "arousal": 0.6 }
        }
    }
}

4.5 Metadata Object

Field	Type	Description
model_version	string	Version of the analysis model
model_name	string	Name of the model used
processing_time_ms	integer	Time taken to process
device_id	string	Identifier for input device
culture_context	string	Cultural context for interpretation
wia_certified	boolean	Whether model is WIA certified

4.6 Complete Example

{
    "format": "WIA-EMOTION-AI-v1.0",
    "timestamp": "2025-12-19T10:30:00.123Z",
    "subject_id": "user_12345",
    "session_id": "sess_20251219_001",
    "modality": "facial",

    "emotions": {
        "primary": {
            "label": "happiness",
            "confidence": 0.87
        },
        "secondary": {
            "label": "surprise",
            "confidence": 0.23
        },
        "all": [
            { "label": "happiness", "confidence": 0.87 },
            { "label": "surprise", "confidence": 0.23 },
            { "label": "neutral", "confidence": 0.15 },
            { "label": "sadness", "confidence": 0.05 },
            { "label": "anger", "confidence": 0.03 },
            { "label": "fear", "confidence": 0.02 },
            { "label": "disgust", "confidence": 0.01 }
        ]
    },

    "dimensions": {
        "valence": 0.72,
        "arousal": 0.45
    },

    "action_units": [
        { "au": "AU6", "intensity": 0.8, "name": "Cheek Raiser" },
        { "au": "AU12", "intensity": 0.9, "name": "Lip Corner Puller" },
        { "au": "AU1", "intensity": 0.3, "name": "Inner Brow Raiser" },
        { "au": "AU2", "intensity": 0.25, "name": "Outer Brow Raiser" },
        { "au": "AU25", "intensity": 0.4, "name": "Lips Part" }
    ],

    "face": {
        "detected": true,
        "count": 1,
        "bbox": { "x": 120, "y": 80, "width": 200, "height": 250 },
        "landmarks": 68,
        "pose": { "pitch": 5.2, "yaw": -3.1, "roll": 1.2 },
        "quality": { "blur": 0.1, "exposure": 0.8, "occlusion": 0.05 }
    },

    "metadata": {
        "model_version": "2.1.0",
        "model_name": "WIA-EmotionNet-v2",
        "processing_time_ms": 45,
        "device_id": "webcam_001",
        "culture_context": "western",
        "wia_certified": true,
        "certification_level": 2
    }
}

4.7 JSON Schema Definition

{
    "$schema": "http://json-schema.org/draft-07/schema#",
    "$id": "https://wiastandards.com/emotion-ai/v1/schema.json",
    "title": "WIA Emotion AI Data Format",
    "type": "object",
    "required": ["format", "timestamp", "modality", "emotions"],
    "properties": {
        "format": {
            "type": "string",
            "pattern": "^WIA-EMOTION-AI-v[0-9]+\\.[0-9]+$"
        },
        "timestamp": {
            "type": "string",
            "format": "date-time"
        },
        "modality": {
            "type": "string",
            "enum": ["facial", "voice", "text", "biosignal", "multimodal"]
        },
        "emotions": {
            "type": "object",
            "required": ["primary"],
            "properties": {
                "primary": { "$ref": "#/definitions/emotionScore" },
                "secondary": { "$ref": "#/definitions/emotionScore" },
                "all": {
                    "type": "array",
                    "items": { "$ref": "#/definitions/emotionScore" }
                }
            }
        },
        "dimensions": {
            "type": "object",
            "properties": {
                "valence": { "type": "number", "minimum": -1, "maximum": 1 },
                "arousal": { "type": "number", "minimum": -1, "maximum": 1 },
                "dominance": { "type": "number", "minimum": -1, "maximum": 1 }
            }
        },
        "action_units": {
            "type": "array",
            "items": { "$ref": "#/definitions/actionUnit" }
        }
    },
    "definitions": {
        "emotionScore": {
            "type": "object",
            "required": ["label", "confidence"],
            "properties": {
                "label": { "type": "string" },
                "confidence": { "type": "number", "minimum": 0, "maximum": 1 }
            }
        },
        "actionUnit": {
            "type": "object",
            "required": ["au", "intensity"],
            "properties": {
                "au": { "type": "string", "pattern": "^AU[0-9]+$" },
                "intensity": { "type": "number", "minimum": 0, "maximum": 1 },
                "name": { "type": "string" }
            }
        }
    }
}

4.8 Chapter Summary

[OK] Key Takeaways:

JSON Format: All emotion data uses JSON with defined schema
Required Fields: format, timestamp, modality, emotions
Emotion Models: Supports both discrete labels and V-A dimensions
Action Units: FACS AU encoding with 0-1 intensity
Modality Extensions: Specific fields for face, voice, text, biosignal
Multimodal: Fusion results with modality weights
Validation: JSON Schema enables automated verification

4.9 Looking Ahead

In Chapter 5, we will explore Phase 2: API Interface, covering REST API endpoints for emotion analysis across all modalities.

Chapter 4 Complete | Approximate pages: 16

Next: Chapter 5 - Phase 2: API Interface

WIA - World Certification Industry Association

Hongik Ingan - Benefit All Humanity

https://wiastandards.com