Today, there are many algorithms for recognition and detection of faces in the image. In addition, each of them has input parameters on which the result strongly depends. The result also depends on the lighting, contrast, angle and turning of the face. Thus, there are a huge number of options for detecting and recognizing faces in an image. Determining the best parameters of algorithms and images with such a huge number of variations is almost impossible manually.
To solve this problem, a software-hardware complex is being developed that will simplify the study of facial recognition algorithms.
In this paper, the means of implementing the software-hardware complex are selected and its architecture is developed. Viola-Jones algorithm was chosen for face detection, as it is the most appropriate in terms of the accuracy / speed of face detection from the video stream. In the work, the algorithms EigenFaces, FisherFaces, LBP, AAM, ASM were chosen as algorithms for studying within a software-hardware complex. Also, development tools are selected: Django, QT, Celery, Redis, PostgreSQL, OpenCV, Python. The client-server architecture of the complex is presented. The server part of the complex includes face recognition modules, drawing graphs and histograms, determining faces on the incoming image, face recognition and its classification. The algorithm of user interaction with the complex is presented: the user, using the client application, sends a face image, the server part of the complex processes the image and classifies it based on the existing people in the database. After processing the image of the face, the user can get the results of classification and vary the parameters of the original image, getting new results.
Also, the work presents the image parameters that affect the accuracy of facial recognition and the developed program interface.
1. Tatarenkov D.A. Analiz metodov obnaruzheniya lits na izobrazhenii [Analysis of face detection methods in the image]. Molodoi uchenyi – Young Scientist, 2015, no. 4 (84), pp. 270–276.
2. Tikhonova T.S., Belov Yu.S. Osnovnye podkhody k otslezhivaniyu i raspoznavaniyu litsa [The basic approaches to face recognition and tracking]. Elektronnyi zhurnal: nauka, tekhnika i obrazovanie – Electronic journal: science, technology and education, 2016, no. 2 (6), pp. 111–115.
3. Razinkin V.B., Katermina T.S. Raspoznavanie litsa po fotografii [Face recognition by photo]. Mezhdunarodnyi zhurnal perspektivnykh issledovanii – International Journal of Advanced Studies, 2018, vol. 8, no. 1-2, pp. 171–180.
4. Analiz sushchestvuyushchikh podkhodov k raspoznavaniyu lits [Analysis of existing facial recognition approaches]. Available at: https://habr.com/ru/company/synesis/blog/238129/ (accessed 27.05.2019).
5. Face recognition with OpenCV. Available at: https://docs.opencv.org/2.4/modules/contrib/doc/facerec/facerec_tutorial.html (accessed 27.05.2019).
6. Django documentation. Available at: https://docs.djangoproject.com/en/1.11/ (accessed 27.05.2019).
7. Samal D.I., Frolov I.I. Algoritm podgotovki obuchayushchei vyborki s ispol'zovaniem 3D-modelirovaniya lits [Algorithm of preparation of the training sample using 3d-face modeling]. Sistemnyi analiz i prikladnaya informatika – System analysis and applied information science, 2016, no. 4, pp. 17–23.
8. Vijayakumari V. Face recognition techniques: a survey. World Journal of Computer Application and Technology, 2014, vol. 1, no. 2, pp. 41–50. DOI: 10.13189/wjcat.2013.010204.
9. Viola P., Jones M.J. Robust realtime face detection. International Journal of Computer Vision, 2014, vol. 57, no. 2, pp. 137–154.
10. Turk M.A., Pentland A.P. Face recognition using eigenfaces. Proceedings. 1991 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Maui, HI, USA, 2017, pp. 586–591. DOI: 10.1109/CVPR.1991.139758.
11. Bui T.T.T., Phan N.H., Spitsyn V.G. Raspoznavanie lits na osnove primeneniya metoda Violy-Dzhonsa, veivlet-preobrazovaniya i metoda glavnykh komponent [Face recognition using ViolaJones method, wavelet transforms and principal component analysis]. Izvestiya Tomskogo politekhnicheskogo universiteta – Bulletin of the Tomsk Polytechnic University, 2012, vol. 320, no. 5, pp. 54–59.
12. Sirovich L., Kirby M.A. Lowdimensional procedure for the characterization of human faces. Journal of the Optical Society of America. A, Optics and image science, 1987, vol. 4, no. 3, pp. 519–524.
13. Kashem M.A., Nasim Akhter, Md., Shamim Ahmed, Mahbub Alam, Md. Face recognition system based on principal component analysis (PCA) with back propagation neural networks (BPNN). International Journal of Scientific & Engineering Research, 2011, vol. 2, iss. 6, pp. 1–10.
14. Barker S.E., Powell H.M. High speed face location at optimal resolution. World Congress on Neural Networks: 1995 International Neural Network Society Annual Meeting Washington, DC, 1995, vol. 2, pp. 536–541.
15. Lin K.H., Guo B., Lam K.-M., Siu W.-C. Human face recognition using a spatially weighted modified Hausdorff distance. Proceedings of 2001 International Symposium on Intelligent Multimedia, Video and Speech Processing. ISIMP 2001, Hong Kong, China, 2001, pp. 477–480.
Il'in A.A. Razrabotka programmno-apparatnogo kompleksa dlya izucheniya algoritmov raspoznavaniya lits [Development of software-hardware complex for the study of face recognition algorithms]. Sbornik nauchnykh trudov Novosibirskogo gosudarstvennogo tekhnicheskogo universiteta – Transaction of scientific papers of the Novosibirsk state technical university, 2019, no. 1 (94), pp. 32–41. DOI: 10.17212/2307-6879-2019-1-32-41.