Computer Vision Solutions

Image by Daniil Kuželev

Why use Deep Blue AI's Computer Vision solutions?

Accuracy.png

Application-specific image processing techniques

Customisable.png

Customised cutting-edge model architectures

Feedbackv2.png

Built-in feedback for continuous improvement

Computer vision makes use of the vast amount of unstructured data in the form of pictures and videos that has been made available in recent years through reductions in cost of CCD and CMOS sensors. 

The latest generation of neural network architectures is able to pick out nearly invisible features from images to perform many tasks such as object detection, facial recognition, and image classification.  

Deep Blue AI has developed a flexible modular platform that allows easy integration of our custom cutting-edge computer vision models into your interface using a set of APIs.

Our computer vision frameworks

Object Detection

ObjectDetection.png

Detect all objects of interest within a single image and locate them within bounding boxes to place an object relative to others.

Image Classification

ImageClassification.png

Automatically classify images of objects into two or more categories to treat them in batches. 

Facial Recognition

FacialRecognition.png

Detect whether a given face is the same as another face to enhance physical security.

Computer vision solutions in detail 

 

Object Detection

ObjectDetection.png

Retina

Detecting, identifying, and placing objects from the perspective of a camera sensor is of utmost importance in any solution that aims to autonomously move machinery within an uncontrolled space.

 

Deep Blue AI has developed Retina, an AI engine to detect objects contained in an image and draw bounding boxes around them. When used with two sensors, Retina can locate these objects with respect to the sensors in all three axes, making it possible to safely, efficiently, and effectively move machinery autonomously. 

Image Classification

ImageClassification.png

Iris

Classifying images into two or more groups has applications across a range of industries. The unstructured nature of images and the huge variability in picture quality and lighting conditions has meant that this task is currently performed by humans.

 

The latest generation of neural network architectures comes to life in Deep Blue AI’s Iris image classification engine. With Iris, you can classify images with incredible accuracy in fractions of a second. Whether you’re differentiating between grades of material samples, different species of plants, or medical images, Iris has the power and the flexibility for your needs. 

 
 

Facial Recognition

FacialRecognition.png

Pixel

Keys, swipe cards, and PIN codes are important parts of physical security in many workplaces today. They are all, however, susceptible to theft and consequent misuse. Biometric identifiers such as fingerprints and facial recognition offer much higher levels of security while making it quicker and more hassle-free for employees to gain access. 

Pixel is Deep Blue AI’s facial recognition engine. Given a database of allowed faces, Pixel can tell if an image is of an allowed face or not. Using Pixel, you can both boost and streamline your security.