Abstract
Background. Today there are dozens of methods and approaches for facial recognition. In recent years there were achieved more or less acceptable results in object localization, but object identification is still an open problem for achieving high reliability of the correct person identification. It depends on collecting initial data very much. Therefore, there is a need for image preprocessing, which will help to improve the reliability of existing image recognition methods and computer vision approaches by manipulating the input signal.Objective. The purpose of the paper is to search and analyze effective image processing techniques based on key points of the face. Performing experiments on the proposed image preprocessing method and comparing the results of experiments with and without such processing.Methods. The purpose was achieved by analyzing the known identification methods that were used in criminalistics for the last decades and also by the image transformation based on the object's key points with the use of the image normalization approach based on the geometric parameters in computer vision issues.Results. There are the results of studies of various methods of learning and recognition in this research. The time values of training and recognition with and without the proposed image preprocessing method are analyzed. It is shown that by using this image preprocessing method it can achieve better results of object identification.Conclusions. The developed face anthropometric points localisation methods differ from the existing ones by the ability to find key features on low-contrast images with a complex background in real time without prior training and adjustment. The developed image preprocessing method allows improving the reliability of existing recognition methods by the image normalization based on the face key features.
Highlights
Today there are dozens of methods and approaches for facial recognition
It depends on collecting initial data
there is a need for image preprocessing
Summary
Today there are dozens of methods and approaches for facial recognition. In recent years there were achieved more or less acceptable results in object localization, but object identification is still an open problem for achieving high reliability of the correct person identification. It depends on collecting initial data very much. There is a need for image preprocessing, which will help to improve the reliability of existing image recognition methods and computer vision approaches by manipulating the input signal
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.