Search academic texts

Information × Registration Number 0309U000056, 0107U012236 , R & D reports Title Face detection improvement using the combined cascade of classifiers. popup.stage_title Покращення виявлення облич на основі комбінованого каскаду класифікаторів Head Sachenko Anatoly Olexiyovych, Registration Date 16-01-2009 Organization Ternopol National Economic University popup.description2 Project objective: to develop method, algorithms and system of the face detection in the videoflow with low false positive alarms rate. Research object: face detection process in the videoflow.Research methods: pattern recognition theory, image processing and mathematical statistics methods, artificial neural networks’ theory. Results of researches: the face detection method is improved using the combined cascade of neural networks-based classifiers, allowing to increase detection reliability; the face-candidates’ verification method is developed first, allowing to increase the combined cascade of neural networks-based classifiers’ processing speed; the convolutional neural network’s (for the face-candidates’ verification) training method is improved, allowing to decrease the trained network’s classification error for positive examples within the combined cascade of neural networks-based classifiers.5635 Product Description popup.authors popup.nrat_date 2020-04-02 Close