RetailTech

keys 1

Global Ai integrator

1. Development of an online fitting room for sofa covers

Technologies used:

Semantic Segmentation, GAN, pix2pix;

Component:

Object Classification

Task:

Embedding a solution on the company's website that will allow users to upload a photo with a sofa and try on a cover from the store on it

Solution:

Development of a neural network to determine the sofa in the photo

Development of a texture mapping algorithm taking into account the depths and shadows of the image

Transfer and adaptation of the solution to servers

Embedding the solution on a website

Result:

Increase of involvement of the website audience in the purchase process

Increase of competitive advantage

Increase of conversion from shopping cart to sale

2. Automated access and video analytics system

Technologies used:

Biometric scanner based on facial and voice identification

Task:

Systematization of the collection and analysis of the emotional loyalty of buyers. To carry out qualitative research of buyers (surveys, interviews).

Solution:

The administrator sets parameters for calculations: what data to collect, what period of data collection, the task for calculation

Collecting data using cameras

Sorting and writing data to the database. Calculations based on the data obtained

The system uploads the report to the administration panel

Result:

Determines the gender identity of the respondent with high accuracy, which allows us to hypothesize about the correlation of gender and product features

Allows to make a comparative analysis of products and determine from which taste of the product, customers often enjoyed

3. Smart Testing - emotion recognition systems online and on the record

Technologies used:

A neural network ensemble consisting of a set of models for classifying the input set of images and audio signals into three main groups

Task:

Based on business objectives, determine emotions at each stage of work to increase KPIs at various stages of the funnel

Solution:

Collection and analysis of information about the product understudy

Defining business goals and business objectives for setting up metrics

Design, design and development of an interactive structure used to install the system

Implementation and configuration of a data collection system: a facial recognition system, a system for determining gender, age and emotions

Result:

Food tasting

HR brand development and team level improvement

Use in education and development

4. Emotion recognition

Technologies used:

Neural network

Task:

Automated collection of feedback from people regarding an event/product

Solution:

A set of neural network algorithms trained to recognize target emotions in real-time based on a video sequence (age, emotions, gender)

Result:

Automatic processing of video footage from cameras, building an emotional map of the event/product

Ready to start something great?