Accedo/Neurolab

NeuroLab is a virtual artificial neural network (ANN) laboratory that facilitates the use of this artificial intelligence technique as a tool to solve problems, generally involving pattern recognition, which can be of both scientific and practical interest.

The result is the creation, training, and validation of an artificial neural network with the purpose of recognizing the learned patterns in the data it will analyze.

For imaging, the SIRNA tool is used, which is capable of generating patterns from the images for the training process. Once the trained neural network is obtained, SIRNA is able to recognize the patterns in new images.

From the input patterns representing a cross-sectional image of an artery, the aim is to recognize the areas of the artery with different percentages of fat.

• An artificial neural network is created, configured, trained, and validated with NeuroLab to recognize the areas of the image with different percentages of fat.

Recognizing the same image with the neural network obtained with NeuroLab. Each color represents a different class.
Recognizing states within a kidney scintigraphy image with a neural network obtained with NeuroLab. 

It is capable of calculating the percentage of function of each kidney and of both kidneys together.