Computer Vision

Computer Vision associated topics involve both “smart” algorithms and bread and butter extensive tracing, testing under variety of lighting conditions. Are we dealing with lower quality video-feed or higher-quality individual frames? Is the data compressed, causing the algorithms to stumble? How to select the right cameras in the high-temperature or contaminated environment? We know the landmines and ways around them. And go above and beyond.

We are experts in providing OpenCV-based scalable solutions on both mobile and desktop platforms under heavy use conditions in critical environments.
Extensive experience with OpenCV implementation of various objects detection, classification and image segmentation tasks using classical machine vision algorithms: SVM and cascade classifiers, optical flow, non-parametric clustering, feature matching and visual words techniques. We also provide OpenCV adaptations of the state-of-the-art deep learning neural-network and autoencoder based models, trained in Google Tensorflow and Caffe frameworks.

Computer vision tasks we successfully handled in recent years include:

  • arbitrary objects’ detection and tracking
  • 3D scene reconstructions
  • ID documents real-time detection, geometric correction and OCR analysis face recognition with liveness verification
  • Image quality enhancement, denoising and superresolution
Send us a note and we’ll get in touch with you