Emotion Recognition AI for Extraction of Concealed Feelings


Infocomm - Artificial Intelligence


This technology offer is a healthcare-grade emotion recognition AI with close to 10 years of R&D and commercial use that can discern emotions, high accuracy, and is capable of extracting concealed feelings. The strength of the algorithm arises from the volume and granularity of data collected, enabling the AI engine to detect and assess a broad range of emotions with precision. There are annotated datasets from various fields of psychology; after extensive training and validation, the algorithm recognises the minutiae scale of emotions for applications across platforms. This technology offer can be used in mental health analysis, vehicle safety, education, unbiased emotion feedback for advertising and public security. It has been deployed for audience emotion measurement, customer satisfaction, human resources and training applications, etc.

The technology owner is keen to out-license this technology and/or work with technology collaborators who can further co-develop this technology. E.g., companies with relevant hardware, and require lightweight emotion recognition AI at the edge to bring new products/services into the market, such as non-intrusive emotion recognition cameras in shopping malls.


The features and specifications of this emotion recognition AI include:

  • Using psychology circumplex model with 2 dimensions for subtle expressions analysis
  • Real-time processing and analysis for emotional states of individuals and large groups 
  • The algorithm accounts for ethnicity and cultural differences with high accuracy especially in negative expressions
  • Thousands of mood detection with accuracy of 93%
  • Deployment methods include: Apps, SDK, Raspberry Pi, Cloud Service

There are various annotated datasets:

  • Cognitive, experimental and social psychology fields are used to train our algorithm to detect emotions of drivers and large groups - on safety, interactive infotainment or malicious intent
  • Clinical, counselling and psychological fields are used to train the algorithm to recognize emotions that can be used to aid detection and intervention of mental health.
  • Behavioural, educational, and human factors psychology fields are used to train the algorithm to detect attention and engagement levels of participants/users in education and advertising research.
  • Cross-cultural, developmental, social and forensic psychology fields are used to train the algorithm to detect emotions of people in crowded places for public safety and experience


The potential applications of the emotion recognition AI include: 

  • Real-time emotion detection for mental health, elderly care via remote telehealth platforms and emotional support health companion devices
  • Adaptive learning content, curriculum developed with emotion data and mood-meter for classrooms or exhibitions
  • Objective aptitude training with enlarged behavioral cues on situational training, online training and cognition skill development
  • Online recruitment psychometric evaluation with candidate emotion analysis
  • Emotion insights into A/B testing/ packaging / POS satisfaction survey or guest services concierge measurement
  • Unbiased genuine emotion feedback for focus group or market research analytics
  • Gather emotion data at any displays and make improvements or create more interactive contents

Market Trends & Opportunities

  • Public safety by adding another layer to security at malls, airports, sports arenas, and other public venues to detect malicious intent or crowd moods
  • Smart cities to generate real-time ‘crowd sentiment maps’ in public places in smart cities for security and safety, comfort and pleasure measurement
  • Tele-consultation on real-time emotional state of patients through video and mobile devices to assist in early-stage intervention of non-motor symptoms such as depression, anxiety, cognitive decline
  • Hyper-connected devices to measure and track emotions, measure and analyse well-being
  • Tele-counselling for analysis of moods to help healthcare professionals in the fields of geriatrics and paediatrics to detect non-motor symptoms such as depression, anxiety, apathy, irritability and pain
  • Health prediction and wellness for therapy evaluation, screening tool to detect symptoms for holistic management, novel therapies and therapeutic strategies
  • Digital Out Of Home (DOOH) for audience and context-aware signage displays that leverage the emotion of audience behaviour
  • Events experience to better understand the impact and reactions to specific elements at an event to inform on how to invest in future events and audiences experience
  • Education with measure of real-time learner emotional responses to educational content
  • Talent recruitment with emotional analysis to do accurate pre-employment screening and assessment
  • Automotive and taxis for detecting emotion of drivers and passengers, integrated into infotainment systems and Advanced Driver Assistance Systems
  • Retail and hospitality to interact and assist shoppers, greet customers at the store entrance or customer service desk
  • IoT for surveillance or enabling robotics with emotional intelligence


Benefits of this technology offer include the following:

  • Gain emotion insights non-intrusively for content, A/B testing and packaging
  • Boost content designed in a practical approach beyond spoken or written feedback (emotion feedback is more genuine)
  • Enhance survey measurement (high survey scores but low emotion recognition, as users tend to be polite when giving feedback)
  • Strengthen market research methods to formulate preferences based on emotions in response to new services or products
  • Deepen behavioral methods and automate emotion analysis to scale data collection efforts from multiple sources.
  • Discover emotions in real-time for customers service, improve productivity, save costs and make user experience 
  • Quantify consumer behavior trends - optimise everything from shelf-level displays to store layout based on customers’ emotion analytics
  • Link emotion analytics data on multiple IoT devices, to have broad understanding of user behaviors
  • Extend emotion detection for better measurement in interviews, employee morale and HR tools on recruiting strategies
  • Humanise computer interactions in intelligent personal assistant, concierge service and customer service
  • Infuse emotion intelligence into devices making them truly social, empathic and smarter
  • Expand DOOH marketing measurement in public places, in transit, commercial, retail venue, billboards, street and roads
  • Widen emotional analysis at point of experience for a greater understanding of behavior patterns to help predict likely future purchasing
  • Insight into what impacts customer emotions to help drive sales, improve product and service offerings and experiences
Autonomous Mobile Robots Navigation Software
Estimated Time of Completion (ETC) Prediction for Last-Mile Logistics
Software Tool to Assess User Attentiveness During Online Engagements
Visual 3D Positioning & Mapping System for Autonomous Vehicles and Robots
Accelerated Materials Innovation Platform for Sustainability
Real Time, All-day, Stress Monitoring System Using Data Science
AI-Based Medical Imaging Assistant For Breast Cancer Screening
Chatbot Training Platform
Improving Explainable Artificial Intelligence For Degraded Images
Enhancing Construction Safety and Productivity with Video Analytics