Emotion recognition

Emotion recognition

This technology has been trained to track the features of a human face and recognize a number of emotional responses. Potential applications include collecting emotional feedback both in controlled environments, like tests, surveys, or focus groups, and out in the retail settings, like promos and window displays.

Features

icon

Uses a device camera to read facial expressions

icon

Neutral expression and +6 base emotions

icon

Trained by analyzing screen activity of many volunteers

icon

Works with both photo and video input

Application Examples

Nicola app

Implemented to collect emotional feedback when reading

Design testing

Collecting emotional feedback when testing interfaces and products

Empathetic interfaces

Games, chats, and other digital products that react to user’s emotions

Promo analysis

Analyzing emotional reactions to both online and in-store promos

Activity recognition

Activity recognition

Activity recognition is a problem of classifying activity that a human performs from available sensor data. By activity we may imply movements, behaviours or cognitive tasks. Our technology solves the problem of PC user activity detection. Using a real time camera video stream we can recognize a number of workloads from reading and quick text skimming to media content consumption.

Features

icon

Uses a device camera to track eye movement. Works on over 90% of all Android and iOS devices

icon

Patent-pending combination of deep learning and automatic pattern detection

icon

Uses neural networks, optimized for mobile devices

icon

One of the most energy efficient eye-tracking solutions for mobile

Application Examples

Nicola app

Implemented to collect emotional feedback when reading

Educational platforms

In conjunction with emotion recognition analyze user’s attitude to a certain type of workloads

Advanced user interface engines

System user interface and behaviour may automatically adapt to the task a user performs

Time-tracking apps

Analyze work and non-work related workloads to generate employee efficiency reports

Prediction of perceived personality traits

Prediction of perceived personality traits

With this technology, we can employ user’s features and facial expressions to predict the impression they will make on other people. We envision the technology being used by public speakers, job seekers, and anyone looking to improve their social interactions.

Features

icon

Uses a device camera to track facial features and expressions. Works with both photo and video input

icon

Predicts openness, conscientiousness, extraversion, and agreeableness

icon

Guarantees impression accuracy of over 80%

icon

Based on the Big Five personality traits theory

Application Examples

FaceMe app

The technology is adopted in our own personality test app

Public speaking

Rehearsing and analyzing the delivery of a speech

Promo optimization

Analyzing facial features and expressions within marketing materials

Job interviews

Preparing for an interview, evaluating job seekers

Real-time Pose Estimation

Real-time Pose Estimation

This technology uses camera to look at a human body and track every movement in real time. We envision it being implemented for anything that requires an augmented reality mirror, be it physical exercise, virtual dressing room, or even Kinect-like gaming.

Features

icon

Uses a device camera to track eye movement

icon

Based on Banuba AR SDK

icon

Works on over 90% of all Android and iOS devices

icon

Computationally efficient solution. Works with 30+ FPS on advanced devices

Application Examples

Health & fitness apps

Learning the correct way to exercise, counting reps

Gesture-controlled interfaces

Using gestures to interact with games and other digital products

Clothes shops

Trying on clothes while online shopping

Nicola app

Implemented to collect emotional feedback when reading

Text complexity & Age restrictions

Text complexity & Age restrictions

Building a content recommendation system like the one inside digital education platform or electronic library. It is designed to find the content (tasks, books) suitable for a certain user. Among other factors, matching content is the one that can be effectively consumed and well understood. What’s more, it does not violate user age restrictions, if any. We’ve made several prediction units that infer relative complexity (based on the number of words, vocabulary, etc.) and age restrictions (based on the text language and expressions used).

Features

icon

Strictly checking a text topic and vocabulary for age

icon

Trained on 10, 000 000+ of dialogues from nearly 200,000 movies and shows

icon

The coefficients of the general word usage are calculated on the basis 57,000 books

icon

Alternative to proprietary Amazon Lexile Measure

Application Examples

Books recommendation system

Smart filtering of the inappropriate or adult content

Nicola app

Recommend books or education tasks based on user age and reading skill

Fatigue & Drowsiness Detection

Fatigue & Drowsiness Detection

Tracking whether the distance between eyelids is narrowing and analyzing the time and frequency of blinking. It can detect fatigue and loss of concentration on the action being performed.

Features

icon

Uses a device camera to track eye movement

icon

Based on Banuba AR SDK

icon

Works on over 90% of all Android and iOS devices

icon

Computationally efficient solution that works with 30+ FPS precision on advanced devices

Application Examples

Nicola app

Implemented to collect emotional feedback when reading

Control over drivers and staff

Accident prevention and rest recommendations

Employee monitoring

Efficiency measurements and optimization of rest time