Train neural networks on 2D images for object detection/classification, noise reduction and segmentation

Go to Teledyne Dalsa
Go to Machine Vision Software
Go to Artificial Intelligence Software

Astrocyte Artificial Intelligence software

 

  
 

Based on artificial intelligence, Astrocyte is an application that is designed to train neural networks on 2D images for a wide variety of applications. With an interface that delivers impressive flexibility, users can use their own image samples to train neural networks for the purpose of unique object detection/classification, noise reduction and segmentation.

Astrocyte offers visualising and interpreting models for optimised performance and reliability. Furthermore, you can export these models for later use at runtime using Teledyne Dalsa's Sapera and Sherlock platforms.

 
 
   
 
Key Features
 
 
• Supports Microsoft Windows (Linux to come)
• Pre-trained models for reduced training effort
• Export of model file to interface with Sapera Processing and Sherlock for runtime inference
• Visualization of model performance with numerical metrics and heatmaps.
• Graphical view of training progress with possibility of cancelling/resuming sessions
 
• Access to hyperparameters for highly flexible training, including selection of neural network type.
• Import of training samples from local or remote locations with various file selection schemes.
• Multiple deep learning architectures for a wide range of applications.
• Training deployed on user PC for full privacy (no need to share data on cloud).
 
     
     
     
     
  Noise Reduction   Segmentation  
   


 

 
  Noise reduction intends to deliver a higher-quality images from its original state. It is an important component in applications such as digital photography, medical image analysis, remote sensing, surveillance and digital entertainment.  

Image segmentation is important component in computer vision and is used for defect sorting/qualification, food sorting, shape analysis, etc. Image segmentation involves dividing input image into segments to simplify image analysis.

 

  
  Object Detection   Object Classification  
     
  Object Detection involves identifying one or more objects of interest in an image. Object Detection is used to solve problems like presence detection, object tracking, defect localization and sorting, etc.   Classification involves predicting which class an item belongs to. Classification is used to solve problems like detect identification, character recognition, presence detection, food sorting, etc.  
 

 

 

  

 

  
 

Both Sapera Processing and Sherlock include an inference tool for each of the supported model architectures. Model files are imported into the inference tool and ready for execution on live video stream. The inference tool can be coupled with other image processing tools such as blob analysis, pattern matching, barcode reading, etc.

The examples provided include source code which can be used conjunction with Sapera LT for acquiring images from Teledyne DALSA cameras and frame-grabbers.

Sapera LT is an image acquisition control software development toolkit for Teledyne Dalsa's cameras and frame grabbers. Hardware independent, Sapera offers a rich development system for machine vision OEMs and system integrators.

  
         
  Need a price or more application information? Please email Adept Turnkey or call our offices
Adept Turnkey Pty Ltd is "The Machine Vision and Imaging Specialists" and distributor of Teledyne Dalsa products in Australia and New Zealand. To find out more about any Teledyne Dalsa product, please call Adept Turnkey at Perth (08) 9242 5411 / Sydney (02) 9905 5551 / Melbourne (03) 9384 1775 or contact us online.