20.500.12556/RUL-158700 Avtomatsko ocenjevanje testa tapkanja s prsti pri bolnikih s Parkinsonovo boleznijo Automatic assessment of finger tapping in patients with Parkinson's disease Parkinsonova bolezen je kronična nevrodegenerativna bolezen, ki močno poslabša kvaliteto življenja pacientov. Čakalne vrste za nevrologa so danes precej dolge, pacienti pa so v tem obdobju brez ustrezne terapije s katero bi si olajšali simptome. Zato smo razvili avtomatsko metodo za določanje stopnje motorične prizadetosti oziroma bradikinezije Parkinsonove bolezni na podlagi testa tapkanja in s tem omogočili hitrejšo postavitev diagnoze. Zbrali smo 183 video posnetkov tapkanja, posnetih kar s pametnim telefonom v vsakdanjem okolju. Videe je v 5 razredov lestvice MDS-UPDRS ocenil nevrolog. Za prepoznavo roke smo uporabili MediaPipe Hand, ki nam kot rezultat vrne časovno vrsto skeleta roke. Za klasifikacijo smo ubrali dva različna pristopa. Prvič smo iz časovne vrste skeleta roke sami sestavili značilke, enkrat strogo po lestvici MDS-UPDRS, drugič pa se te nismo strogo držali. Te značilke smo nato uporabili v klasifikatorjih in z večplastnim perceptronom dosegli 61 \% točnost in 0,62 F1 vrednost. V drugem pristopu smo časovno vrsto razdalj med palcem in kazalcem uporabili neposredno v polnem konvolucijskem nevronskem omrežju in dosegli 77 \% točnost in 0,75 F1 vrednost. Izdelali smo še orodje za vizualizacijo tapkanja in izpis ključnih podatkov. Parkinson's disease is a chronic neurodegenerative disorder that severely impairs patients' quality of life. Currently, the waiting lists for a neurologist are quite long, and during this period, patients are without adequate therapy to alleviate their symptoms. Therefore, we have developed an automatic method to determine the level of motor impairment or bradykinesia of Parkinson's disease based on a tapping test, thus enabling faster diagnosis. We collected 183 tapping videos recorded with a smartphone in everyday environments, which were assessed by a neurologist into 5 classes of the MDS-UPDRS scale. For hand detection we used MediaPipe Hand, which returns a time series of the hand skeleton. For classification, we took two different approaches. First, we constructed features from the hand skeleton time series, once strictly following the MDS-UPDRS scale, and another time not strictly adhering to it. These features were then used in classifiers and achieved 61 \% accuracy and 0,62 F1 score using a multi-layer perceptron. In the second approach, we used the time series of thumb-pointer distances directly in a fully convolutional neural network achieving 77 \% accuracy and 0,75 F1 score. We also created a tool for visualizing tapping and displaying key data. klasifikacija Parkinsonova bolezen test tapkanja s prsti strojno učenje classification Parkinson’s disease finger tapping test machine learning true false false Slovenski jezik Angleški jezik Magistrsko delo/naloga 2024-06-19 13:35:01 2024-06-19 13:35:06 2024-07-12 09:55:17 0000-00-00 00:00:00 2024 0 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 37143 201613827 Zupanic_Matjaz_-_Avtomatsko_ocenjevanje_tapkanja_s_prsti_pri_bolnikih_s_Parkinsonovo_boleznijo.pdf Zupanic_Matjaz_-_Avtomatsko_ocenjevanje_tapkanja_s_prsti_pri_bolnikih_s_Parkinsonovo_boleznijo.pdf 1 278B9B60135C969B39E6530E2A57F104 c68b3963e6d980f1045820d93c1bf759f0cb09a29fd506ec0f78ec3451fabcfb fb1a5d70-2e2f-11ef-aa1b-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=186413 Fakulteta za računalništvo in informatiko 0 0 0