Moscow

Connectome.ai

Etalon.ai automates product quality control on assembly lines. It visually examines each product, makes an assessment and notifies the employee in case it detects any defect. The system is self-learning, it uses developments in the field of computer vision and video analytics.&nbsp;

The system includes &nbsp; a set of high-resolution cameras and a video analytics processing module.

Smart system of visual quality control on production lines

video_surveillance

video_analytics

computer_vision

artificial_intelligence

Etalon.ai

Medsenger

Medsenger.AI Platform is the extension of the functionalities of&nbsp;<a href="https://ict.moscow/card/medsenger/" target="_blank">Medsenger</a>. The platform makes it possible to connect remote dialogue "patient-doctor" with specialized intelligent agents, which automatically (but under the supervision of the attending physician) implement a certain set of functions that are relevant specifically for this patient.&lt;

The service is a cloud platform on the basis of which medical organizations offer their patients a new chargeable service: remote consultation (monitoring, support) of a patient by his attending physician between face-to-face visits. The service Canis available on &nbsp; Web, Android, iOS platforms.

The main principle of the service is asynchronous: the patient asks when there is a question, the doctor answers when there is an opportunity. Some functions of remote patient management using intelligent agents (AI) are automated to facilitate the work of the doctor, in particular:

<ul>
	<li>COVID-19 symptoms monitoring, which is free for all users. The service sends a specially designed questionnaire with possible symptoms of coronavirus to the patient according to the doctor’s schedule. In case the patient notices any of the symptoms, the doctor will be immediately notified about this and contact the patient.</li>
	<li>Monitoring of medical parameters and drug intake, which makes it possible to monitor the changes in the patient's condition while taking drugs, assess the effectiveness of the treatment and consider its adjustment, with automatic generation of schedule based on the parameters, which is available for both the doctor and the patient in personal account.</li>
</ul>

Platform for connecting AI tools to remote patient consultations

telemedicine

covidtech

cloud_service

Medsenger.AI

Center2M

CenterVision uses neural networks for object recognition. The system recognizes objects with an accuracy of 98%, track production process in real time, transmit &nbsp; automatic notification to the situation center, prepare &nbsp; reports on security violations.

The system consists of video surveillance cameras, rangefinders, lasers, notification and information systems, software (model training, business logic, scenario adaptation), systems for integration with external scenarios (API), and a user interface.

The response time of 2.4 seconds allows the system to learn about security violations; objects are detected at large distances, regardless of rain, snow or fog, recognition of small objects from a long distance (50-70 m) makes it possible to use fewer cameras for monitoring.

CenterVision

Beorg

Smart Vision provides recognition of documents within 2 minutes with a quality assurance of more than 99%. Types of documents: passports with registration, including handwritten; sets of documents of legal entities and individual entrepreneurs; working documents and personal data of employees; sets for registration of bank cards. Scan formats: image files: pdf, png, tiff, jpg, receiving documents via API.

The service is useful for banks - when registering new clients, insurance companies - when applying for insurance, credit organizations - when issuing loans or plastic cards (bonus, discount, credit), mobile applications services - where documents are required to prove your identity.

Document recognition and image digitization service

neural_networks

Beorg Smart Vision

Neurocore

Module based on computer vision technologies using video surveillance cameras allows to automate control over an object and optimize the overall process of allowing vehicles to enter a territory with a limited access level. Various neural network architectures were selected and tested during the development process, and the optimal one was developed.

The number of images with objects that were memorized by the system is about 1 million. The system operates with both CPU and GPU. High recognition accuracy is achieved by the algorithm of developing and implementing the solution: over 70% of the training cycles are conducted in a "test environment". The current solution works using the client’s API, integration is possible in more complex conditions.

Vehicle license plate number and class recognition system

transportation

Neurocore: car recognition

The system operates with different lighting, camera quality and location of objects. More than 10 thousand &nbsp; images of people of different sex and age were used while creating the module.

It works with various types of computing power: both processors and video cards. As of April 2020, the system operates using API connection.

biometry

Neurocore: facial recognition

Intelligence_Retail

Solution for FMCG (Fast Moving Consumer Goods, daily commodities) manufacturers and trade networks. It allows to assess the quality of &nbsp; presence &nbsp; goods &nbsp; at sales points.

A photo of a shelf with goods taken by an employee on a smartphone or tablet is transmitted to the cloud for recognition and analysis. Within 10 seconds, the staff member receives detailed reports on the quality of the layout, which is also available to the management team on the analytics portal.

retail

Intelligence Retail

VoxiAi

VoxiAi is a service for call centers that improves the quality of voice robots by distributing calls among the robot and/or live operators.

With the help of predictive analysis of the conversation history with the client, VoxiAi selects the suitable specialist, transferring a call before or during conversation without changing the voice; determines the situations in which robot could handle the conversation and when it is necessary to engage a live specialist; analyzes how effectively the operator communicates, choosing the most suitable operator for the client in the future; reduces service time by up to 30%.

chatbots

voice_recognition

machine_learning

Biometric_Labs

Voice biometrics and emotion recognition system for the identification of the caller, combating fraud using telephony and connections, determination of a subscriber’s emotional state by voice, automation of meeting record keeping, ASR.

The system can operate in text-dependent (a person pronounces certain words/figures), text-independent (recognition does not depend on text) modes and using figures.

Acoustic signal analysis technologies: voice biometrics, acoustic sensors, meeting recording systems

Biometriclabs

Mechanica_AI

These are virtual&nbsp;sensors (analyzers) based on artificial intelligence technologies. The solution &nbsp; provides an assessment of the important parameters of the technological &nbsp; process without delay, as well as predicts changes in parameters.

The key element of the solution is a machine learning model created using a large amount of historical data. Due to the available parameters of technological process, model determines relevant values of the parameters that are measured by the virtual sensor.

Virtual sensors with machine learning technology

Mechanica Sense

Neuroscanner

The mathematical models and methods system performs automatic detection of defectograms and digitized signals in real time, with proposals for priority, planned and major maintenance, with a built-in system of automatic predictive diagnostics, in neural networks without human intervention with an accuracy of up to 98% (current accuracy rate - 85%)..

The company is a Skolkovo resident.

System for automatic detection of hidden defects using AI methods

Tech Solutions