Computationally Efficient Video Face Recognition Algorithms Based on Frame Selection and Quality Assessment

Student: Kharchevnikova Angelina

Faculty: Faculty of Informatics, Mathematics, and Computer Science (HSE Nizhny Novgorod)

Educational Programme: Data Mining (Master)

Final Grade: 9

Year of Graduation: 2020

The paper considers the problem of increasing the efficiency of facial identification algorithms by video. We propose an approach based on the key video frames selection using various techniques for assessing frame quality. The experimental comparison of the traditional quality assessment methods based on Brightness, Contrast, as well as deep learning technology are considered. We trained several lightweight convolutional neural networks using the methods of fine tuning and knowledge distillation in order to increase the efficiency of the key frame selection stage. The proposed algorithms are compared with the traditional face recognition approach for each frame, and with the well-known clustering methods.

Full text (added May 27, 2020)

Student Theses at HSE must be completed in accordance with the University Rules and regulations specified by each educational programme.

Summaries of all theses must be published and made freely available on the HSE website.

The full text of a thesis can be published in open access on the HSE website only if the authoring student (copyright holder) agrees, or, if the thesis was written by a team of students, if all the co-authors (copyright holders) agree. After a thesis is published on the HSE website, it obtains the status of an online publication.

Student theses are objects of copyright and their use is subject to limitations in accordance with the Russian Federation’s law on intellectual property.

In the event that a thesis is quoted or otherwise used, reference to the author’s name and the source of quotation is required.

Search all student theses