EngEdu2020
5th Annual International Conference on Engineering Education & Teaching, Athens, Greece, 1-4 June 2020 ...

EGU 2020
European Geosciences Union (EGU) General Assembly, Vienna, Austria, 4-8 May 2020 ...

All news

Publications

 

  Follow us on Geophysical Applications Processing srl on FacebookGeophysical Applications Processing srl on TwitterGeophysical Applications Processing srl on LinkedInGeophysical Applications Processing srl on YouTube



The SINACH research project

Medical GAP Software Lab - Sinach Project

The segmentation of magnetic resonance images represents a fundamental step in various medical fields, including integrated navigation systems for invasive surgery, which are the subject of SINACH, a research project funded by the Apulia region as part of the INNONETWORK 2017 program.


The project involves Italian companies and research centers at the forefront in the medical and image processing sectors.

In recent years, there has been an increasing use of diagnostic techniques based on the segmentation of magnetic resonance images.

In this scenario, it is increasingly important to have segmentation results very quickly and free of artefacts, as much as possible.

In this context, the GAP Software Lab developed by GAP as part of the SINACH project, is a solution able to satisfy both the operator who needs reliable results in a very short time, and the most demanding operator, who wishes to experiment with different segmentation techniques, such as those based on active contour methods, or experimenting the combination of several segmentation techniques, starting for example from thresholding techniques, and then refining them with active contour techniques. In more complex cases, it is also possible to operate manually with a whole series of manual refinement tools that allow the operator to act even at level of the single slice to be segmented.


The GAP Software Lab, developed as part of the SINACH project, therefore offers the operator a real software laboratory in order to experiment with original and state-of-the-art approaches.

The beating heart of this software is based on artificial intelligence algorithms using advanced machine learning techniques capable of generating reliable results through a fully automatic procedure that does not require any intervention by the operator.

The cases to be analyzed are all entered in a single working directory, and with a simple click, the machine learning algorithm will process all the data in this directory, saving the results in the output directory, where the results of the segmentation will be available for each patient.

It is possible to analyze the results of the segmentation produced by the machine learning algorithm slice by slice.

Where the segmentation result does not satisfy the operator, it is always possible to apply refinements based on state-of-the-art algorithms. The refinements can be applied to the single slice or even to the entire data stack, with the possibility of accessing and changing all configuration parameters of these algorithms, until the desired result is obtained.

During these refinement phases, it is possible to accept the result obtained, but also to restore the initial segmentation.

Once the desired quality level has been reached, the segmentation result can be saved in standard formats, easily viewable with all the main software used in the medical sector.

Furthermore, the result obtained through the refinement operations does not represent only an improved version of the product generated by the machine learning algorithm, but in turn can be supplied to the learning algorithm, in order to improve its future performance and therefore guarantee very high quality levels!

The video of this page is available at the following link: https://youtu.be/zw5-xaRW83Y



 


© GAP SrL - All rights reserved - VAT 06424840723

Powered by BOBONET - Webmaster Davide Oscar Nitti - Cookies Policy