In the closing event of the 11th edition of the Fraunhofer Portugal Challenge - which took place yesterday, October 28th – innovative solutions applied to health were highlighted and won the podium in both categories. Inês Machado, from the Instituto Superior Técnico, and Isabel Oliveira, from the Faculty of Engineering of the University of Porto (FEUP), won the first prizes in the PhD and Master's category, respectively.
With a system similar to the GPS system, but applied to neurosurgery, Inês Machado won first place in the PhD category with Sneaking into the brain with a new GPS-like technology, a technology already validated by neurosurgeons, radiologists and engineers at MIT, Harvard Medical School and Brigham and Women's Hospital. The presented technology leads to a more complete removal of pathologies and helps to avoid important damage to neural structures, which results in a decrease in the patient's morbidity and mortality, a decrease in the cost of health and an improvement in the maximum preservation of neurocognitive functions after the patient survival.
In the MSc category, the 1st prize was awarded to Isabel Oliveira for her idea Understanding the decisions of CNNs: An in-model approach, which proposes the creation of a system, a neuronal network, capable of detecting cancer through images. The results obtained in the detection of visual characteristics associated with cancerous cases are remarkable, but the work now awarded goes further, ensuring that the model used does not remain in a black box, that it is understandable to the specialists involved in the evaluation process. The work focuses on the development of a tool capable of producing visual explanations for the classifier's decisions, highlighting the regions most affected and thus allowing to ensure and improve the process of human evaluation of the results of the algorithms.
1st Prize - Inês Machado (IST) | Sneaking into the brain with a new GPS-like technology
2nd Prize - Renato Panda (Universidade de Coimbra) | Improving music discovery and retrieval with emotionally relevant audio features
3rd Prize - Daniel Dinis (Universidade de Aveiro) | All-Digital Antenna Array Transmitter
1st Prize - Isabel Oliveira (FEUP) | Understanding the decisions of CNNs: An in-model approach
2nd Prize - Pedro Dias (FEUP) | Understanding deep learning decisions in EEG signals
3rd Prize - Bruno Degardin (Universidade da Beira Interior) | Weakly and Partially Supervised Learning Frameworks for Anomaly Detection
The winners of the idea contest were voted by a jury and a panel of experts in the field of technologies, in an event that this year, exceptionally, took place remotely. Organized since 2010, the Fraunhofer Portugal Challenge seeks to encourage cooperation between industry and the scientific community, motivating and rewarding research of practical utility by awarding a prize to students and researchers who best contribute to the philosophy that underpins the Fraunhofer's vision AICOS: create technological solutions with a positive impact on people's lives, as well as its mission to create 'Remarkable Technology / Easy to Use' (Remarkable Technology / Easy to Use). The contest involves two categories, MSc and PhD, with three finalists being awarded in each of them (six in total). The global value of scientific prizes is 9 thousand euros.