Download our Guide How to Build a Marketplace in 10 weeks / 🇺🇦 Stand with Ukraine donate x

Computer Vision

Computer Vision MLOps Platform

Sports, Retail, Construction

SentiFactory is a no-code ML deployment solution which simplifies rollout & monitoring of computer vision Machine Learning models to servers and edge devices. 

What we did

  • Created a collection of pre-build computer vision models and supports all key scenarios like Object Detection, Image Classification, Image Segmentation, Pose Detection, Pose Estimation, Text recognition
  • Made integration which supports import of existing models from applications like Microsoft Lobe to provide true end-to-end no-code ML development experience
  • Provides API-based server deployment options for DigitalOcean cloud provider that relieve the burden of manually configuring the server which removes from the burden of manual server setup
  • Easy deployment of models to edge devices, such as Raspberry Pi and mobile phones (Android)
  • Monitoring of deployed models
  • Redeployment - you can retrain the model with new images or videos and update your model to a new version
  • Custom developed models:

    • Sports - AI-Powered Personalized Coaching, which can estimate how good an exercise is done and count and record excercises and moves.
    • Retail Visual Search - product detection through searching and processing of product data on customer request
    • Construction Safety & PPE Detection - videosurveillance operators can detect abnormal activity or presence or absence of key actors or objects. 

The Results

This solution significantly accelerated the development of custom Computer Vision models. With this platform you can easily create standard cases and deploy, it completely eliminates all manual work from model development. You can easily create all existing types of Computer Vision models Object Detection, Image Classification, Image Segmentation, Pose Detection, Pose Estimation, Text recognition and deploy them on a convenient place for you - server, edge or mobile phone on Android.

Technologies used

  • Python
  • Docker
  • Apache Airflow
  • TensorFlow
  • Ansible

Take a look at our other projects

Startup, IoT, ERP, Multitenancy, Mobile App, SaaS

IoT Solution for Sport Gyms

healthcare, Mobile App, CRM

Healthcare Mobile App for Truck Drivers