10.14489/vkit.2014.05.pp.008-013

Горбацевич В. С.
ИСПОЛЬЗОВАНИЕ ЭЛЕМЕНТАРНЫХ БИНАРНЫХ КЛАССИФИКАТОРОВ ДЛЯ ПОСТРОЕНИЯ БИОМЕТРИЧЕСКИХ ШАБЛОНОВ В ЗАДАЧЕ РАСПОЗНАВАНИЯ ЛИЦ
(с. 8–13)

Аннотация. Рассмотрена задача построения компактных дескрипторов изображений для задач распознавания лиц. В качестве дескриптора предложено использовать бинарный вектор, формируемый набором бинарных классификаторов определенного вида. Дано описание алгоритма обучения таких классификаторов, а также алгоритма отбора наиболее информативных бинарных классификаторов для формирования наиболее компактного дескриптора.

Ключевые слова: локальные бинарные шаблоны; распознавание лиц; машинное обучение; адаптивный бустинг; вейвлеты.

Gorbatsevich V. S.
ELEMENTAL BINARY CLASSIFIERS FOR BIOMETRIC TEMPLATE CONSTRUCTION IN FACE RECOGNITION
(pp. 8–13)

Abstract. Face identification and verification problem is a widely studied problem in technical vision. In this paper, we consider the problem of fast face recognition with small template size. In addition, we describe original technique for building very small biometric templates with high recognition quality. As face template, we use small binary vector, every bit of which is the output of the simple binary classifier. That means we can use very fast and robust Hamming distance for template comparison. As binary classifiers, we use special LBP-like (Local Binary Pattern) classifiers with threshold. We introduce original Boosting-like algorithm for learning and selecting most effective set of these classifiers. For better recognition performance, we work in Cohen-Daubechies-Feauveau-wavelet space. For experiments, we use most popular dataset – FERET (The Facial Recognition Technology) database. Experimental results show that our algorithm achieves better recognition quality/template size performance than the most popular representations. Template extraction and comparison is extremely fast: lesser than 0.05 s (on Core i5-2320 CPU in one thread) for extraction (inc. face detection and normalization), ~10 752 658 templates per second (on Core i5-2320 CPU in one thread) for matching. Our templates also are very small ~600 bytes. That means this technique can be used for building real-time biometric identification systems, embedded biometric systems or as first step identification or verification in more accurate biometric systems.

Keywords: Local binary patterns; Face recognition; Machine learning; Adaptive boosting; Wavelets.

+ - Информация об авторах (About the Authors) Click to collapse

Рус

В. С. Горбацевич (ФГУП «Государственный научно-исследовательский институт авиационных систем» ГНЦ РФ, Москва) E-mail: Этот e-mail адрес защищен от спам-ботов, для его просмотра у Вас должен быть включен Javascript

Eng

V. S. Gorbatsevich (State Research Institute of Aviation Systems State Scientific Center of Russian Federation, Moscow) E-mail: Этот e-mail адрес защищен от спам-ботов, для его просмотра у Вас должен быть включен Javascript

+ - Библиографический список (References) Click to collapse

Рус

1. Cohen A., Daubechies I., Feauveau J.-C. Biortho¬gonal Bases of Compactly Supported Wavelets // Communi-cations on Pure and Applied Mathematics. 1992. V. 45, № 5. P. 485 – 560.
2. Human Detection Using Wavelet-Based CS-LBP and a Cascade of Random Forest / D. Y. Kim et al. // Proc. IEEE Conference on Multimedia and Expo. Melbourne, Australia. 2012. P. 362 – 367.
3. The Facial Recognition Technology (FERET) Database. URL: http://www.itl.nist.gov/iad/humanid/feret/ feret_master.html (дата обращения: 22.03.2014).
4. Ahonen T., Hadid A., Pietikainen M. Face Recog¬nition with Local Binary Patterns // 8th European Conference on Computer Vision (ECCV). 2004. V. 1. P. 469 – 481.
5. Local Gabor Binary Pattern Histogram Sequence (LGBPHS): A Novel Non-Statistical Model for Face Repre-sentation and Recognition / W. Zhang et al. // Proc. of IEEE International Conference on Computer Vision. 2005. V. 1. P. 786 – 791.
6. Histogram of Gabor Phase Patterns (HGPP): A Novel Object Representation Approach for Face Recognition / B. Zhang et al. // IEEE Trans. on Image Processing. 2007. V. 16, № 1. P. 57 – 68.
7. Meyers E., Wolf L. Using Biologically Inspired Features for Face Processing // Int. J. Comput Vision. 2008. № 76. P. 93 – 104.

Eng

1. Cohen A., Daubechies I., Feauveau J.-C. (1992). Biorthogonal bases of compactly supported wavelets. Communications on Pure and Applied Mathematics, 45(5), pp. 485-560.
2. Kim D. Y. et al. (2012). Human Detection Using Wavelet-Based CS-LBP and a Cascade of Random Forest. Proc. IEEE Conference on Multimedia and Expo. Melbourne, Australia, pp. 362 – 367.
3. The Facial Recognition Technology (FERET) Database. Available at: http://www.itl.nist.gov/iad/humanid/feret/feret_master.html (Accessed: 22.03.2014).
4. Ahonen T., Hadid A., Pietikainen M. (2004). Face Recognition with Local Binary Patterns. 8th European Conference on Computer Vision (ECCV). Vol. 1, pp. 469 – 481.
5. Zhang W. et al. (2005). Local Gabor Binary Pattern Histogram Sequence (LGBPHS): A Novel Non-Statistical Model for Face Representation and Recognition. Proc. of IEEE International Conference on Computer Vision. Vol. 1, pp. 786 – 791.
6. Zhang B. et al. (2007). Histogram of Gabor phase patterns (HGPP): a novel object representation approach for face recognition. IEEE Trans. on Image Processing, 16(1), pp. 57-68. doi: 10.1109/TIP.2006.884956
7. Meyers E., Wolf L. (2008). Using biologically inspired features for face processing. Int. J. Comput Vision, 76, pp. 93-104. doi: 10.1007/s11263-007-0058-8.

+ - Заказать электронную версию статьи (Purchase digital version of a single article) Click to collapse

Рус

Статью можно приобрести в электронном виде (PDF формат).

Стоимость статьи 250 руб. (в том числе НДС 18%). После оформления заказа, в течение нескольких дней, на указанный вами e-mail придут счет и квитанция для оплаты в банке.

После поступления денег на счет издательства, вам будет выслан электронный вариант статьи.

Для заказа статьи заполните форму:

{jform=1,doi=10.14489/vkit.2014.05.pp.008-013}

Eng

This article is available in electronic format (PDF).

The cost of a single article is 250 rubles. (including VAT 18%). After you place an order within a few days, you will receive following documents to your specified e-mail: account on payment and receipt to pay in the bank.

After depositing your payment on our bank account we send you file of the article by e-mail.

To order articles please fill out the form below:

{jform=2,doi=10.14489/vkit.2014.05.pp.008-013}