The consortium between the MICLAB Laboratory (Medical Image Computing Lab), from the Faculty of Electrical and Computer Engineering (FEEC) at Unicamp and Neuralmind, a daughter company of Unicamp, was one of those selected for the IdeiaGov technological challenge, which proposes the development of solutions for diagnosing Covid-19, by through artificial intelligence in computed tomography and x-rays.
The presented solution can be used in both chest radiography and computed tomography. What is used in artificial intelligence is called convolutional neural networks (CNN), formulated from biological inspirations and which promote improvements in care and diagnoses.
In this proposed challenge, companies could be approved in two categories: Radiography and Computed Tomography. In total, 15 initiatives were registered, three were approved and the consortium was approved in both categories. For Professor Roberto Lotufo, from Neuralmind, the partnership between the company and the laboratory was essential for this double selection:
“When I read about the notice, I thought we had a chance of winning because we have been working on this for a long time. However, it would not be possible to win this challenge with Neuralmind alone, the connection between our expertise and that of MICLAB was essential for us to be selected”, celebrated Lotufo.
In addition to holding a management position at the company, Lotufo is also a collaborating professor at FEEC, where he created the laboratory at FEEC together with professor Letícia Rittner. The laboratory's activities are focused on processing medical images and according to the professor, they have been developing research to combat Covid-19 since April, which helped participation in the challenge.
“During the pandemic, the students who make up the laboratory were very engaged in studying articles and datasets from around the world related to the disease. This initiative helped us focus on Covid. We have been studying for 4 months how our experience with magnetic resonance imaging applied to brain imaging can leverage new solutions in lung computed tomography imaging,” explained Rittner.
For the executive director of Inova Unicamp, Newton Frateschi, the selection of the consortium is a great example of how companies can work together with the University and use knowledge to boost the social environment. “It is important to highlight that this partnership was selected in the x-ray and tomography categories. This recognition only occurred because the laboratory's competence was seen in relation to the company's innovative capacity, both coming from Unicamp. It is an example of how knowledge can be transformed into innovation and can have an impact, both during the pandemic and in other scenarios”, expressed Frateschi.
About IdeiaGov and next steps
IdeiaGov is an innovation hub that aims to bring market and societal solutions to challenges facing the Government of the State of São Paulo. They are responsible for making the connection between companies and public managers in the development, implementation and scaling of technological and innovative solutions to the proposed challenges. For Lotufo, the purpose is to encourage and promote the development of solutions in the state or country, so that the Government of São Paulo does not depend on international solutions.
The challenge that the consortium was selected for is part of one of the four initiatives proposed to combat coronavirus. Therefore, it selected initiatives with the competence to develop diagnostic solutions, but which, according to Rittner, also aim to actually reach hospitals.
To develop the solution, the selected initiatives will have support from the technical teams of the Hospital das Clínicas of the Faculty of Medicine of the University of São Paulo (HCFMUSP), the Institute of Technological Research of the State of São Paulo (IPT) and the Data Processing Company of São Paulo (PRODESP), in the execution of proofs of concept, pilot tests and non-confidential technical information related to the fight against Covid-19.
Published article originally on the Unicamp Innovation Agency website.