Livoa LogoLivoa
AI-DRIVEN EMPLOYEE MULTIMODAL EMOTION DETECTION SYSTEM
Input Layer


Employee messages / feedback

Preprocessing Layer


Text: Tokenization, Lemmatization

Audio: Noise Removal, Normalization

Image: Face Detection, Resizing

Input Processing


Waveform Analysis

Image Frames

Audio & Text Input


Text Processing

Audio Preprocessing

Image Preprocessing

Result Display


92% Happy

Confidence Score ✓

Remedies & HR Insights

Multimodal Fusion Layer


• Feature-Level Fusion

• Combined Representation

Feature Extraction


Facial & Audio Features

Sequence Learning


Bidirectional LSTM

Classification Layer


Detected Emotions:

• Happy / Sad

• Angry / Neutral

• Fear / Surprise

Emotion Logs Storage


PostgreSQL DB

SYSTEM ARCHTECTURE

by mk

0
0 uses