Wyniki wyszukiwania - BazTech

Ograniczanie wyników

Znaleziono wyników: 2

Liczba wyników na stronie

Wyniki wyszukiwania

Wyszukiwano:
w słowach kluczowych: sieci konwolucyjne

Sortuj według:

Ogranicz wyniki do:

Deep learning vs feature engineering in the assessment of voice signals for diagnosis in Parkinson’s disease

Majda-Zdancewicz Ewelina, Potulska-Chromik Anna, Jakubowski Jacek, Nojszewska Monika, Kostera-Pruszczyk Anna

Bulletin of the Polish Academy of Sciences. Technical Sciences

2021

Vol. 69, nr 3

art. no. e137347

Voice acoustic analysis can be a valuable and objective tool supporting the diagnosis of many neurodegenerative diseases, especially in times of distant medical examination during the pandemic. The article compares the application of selected signal processing methods and machine learning algorithms for the taxonomy of acquired speech signals representing the vowel a with prolonged phonation in patients with Parkinson’s disease and healthy subjects. The study was conducted using three different feature engineering techniques for the generation of speech signal features as well as the deep learning approach based on the processing of images involving spectrograms of different time and frequency resolutions. The research utilized real recordings acquired in the Department of Neurology at the Medical University of Warsaw, Poland. The discriminatory ability of feature vectors was evaluated using the SVM technique. The spectrograms were processed by the popular AlexNet convolutional neural network adopted to the binary classification task according to the strategy of transfer learning. The results of numerical experiments have shown different efficiencies of the examined approaches; however, the sensitivity of the best test based on the selected features proposed with respect to biological grounds of voice articulation reached the value of 97% with the specificity no worse than 93%. The results could be further slightly improved thanks to the combination of the selected deep learning and feature engineering algorithms in one stacked ensemble model.

System wspomagający rozpoznawanie znaków języka migowego oparty na sztucznej sieci neuronowej

Lewandowski P., Półtorak M., Wagner M., Pochmara J., Rybarczyk A.

Poznan University of Technology Academic Journals. Electrical Engineering

2017

No. 91

155--164

W niniejszym artykule zaproponowano realizację systemu wspomagajacego rozpoznawanie statycznych znaków języka migowego. Na potrzeby rozwiązania skorzystano z sensora Microsoft Kinect XBOX 360, przygotowano oprogramowanie umożliwiające translację znaków dla osób nie znających tego języka, oparte na sztucznej inteligencji, przetworzono otrzymane informacje oraz utworzono zbiór danych pozwalający na ich poprawną klasyfikację. Istotnym faktem jest również wybranie najbardziej optymalnego rozwiązania, zarówno pod względem możliwości wydajnościowych przeciętnego komputera osobistego jak i efektywności działania systemu.

In following work there is suggested a solution to recognise certain static characters from sign language. To achieve the objective, there were used tools like Microsoft Kinect and convolutional neural networks. Main problems to overcome were to collect data from Kinect sensor and prepare software based on artificial intelligence, which could process gathered material. For learning purposes around four thousand images were collected. Dataset this large was required for neural networks to work and respond properly. What is also important is to select the most optimal solution for neural networks. The influence of dropout parameter on learning process was studied too.