Detekcija umora vozača iz video snimki

master's thesis

• Title (english): Driver Fatigue Monitoring from Video Data
• Author: Tarik Omanović
• Mentor: Sven Lončarić (mentor)
• Committee member: Sven Lončarić (predsjednik povjerenstva)
• Committee member: Marko Subašić (član povjerenstva)
• Committee member: Tomislav Petković (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2017-07-14, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Vožnja u umornom stanju predstavlja veliku opasnost, stoga postoji realna potreba za detekcijom umora vozača. Detekcija umora iz video signala predstavlja primjenjiv neinvazivan pristup tom problemu. Video signal analiziran je uz pomoć komercijalnog alata FaceTrack proizvođača Visage Technologies te su dobivene značajke lica podijeljene u tri kategorije: jednostavne, složene i one definirane vremenom. Napravljene su video snimke umornih vozača, te je na njima ručno očitana razina umora u svakom... More trenutku. Na video snimkama provedeno je četiri algoritma strojnog učenja. To su k-NN (k najbližih susjeda), SVM (stroj potpornih vektora), AdaBoost i Random forest (nasumična šuma). Svi algoritmi su trenirani na svim skupovima značajki, te su dobiveni ohrabrujući rezultati. Less

Abstract (english)

Drowsy driving poses a threat to public safety, so there is a real need for drowsiness detection. Drowsiness detection from video signal presents a practical, non intrusive approach to this issue. Video signal is analyzed using commercial tool FaceTrack from Visage Technologies. Calculated facial features are divided in three categories: simple features, complex features and features defined by time. Several drowsy individuals have been recorded and level of drowsiness has been manually marked... More on each of them. Four machine learning algorithms have been trained on video signal. Those are k nearest neighbors, SVM, AdaBoost and Random forest. All algorithms have been tested on all sets of features with encouraging results. Less

• Keywords: umor; infracrveno; značajke lica; strojno učenje; konfuzijska matrica; preciznost
• Keywords (english): drowsiness; infrared; facial features; machine learning; confusion matrix; accuracy
• Language: croatian
• URN:NBN: urn:nbn:hr:168:371455
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-03-13 23:29:59