?
Статистическое распознавание образов на основе вероятностной нейронной сети с проверкой однородности
Искусственный интеллект и принятие решений. 2013. № 4. С. 45-56.
Statistical pattern recognition was reduced to the hypothesis test for homogeneity. The probabilistic neural network (PNN) modification was proposed to achieve its optimal decision in terms of minimum Bayes-risk. The comparative analysis' results of the proposed modification with an original PNN were presented in a problem of automatic author identification
Keywords: statistical pattern recognitionprobabilistic neural networkhypothesis test for samples homogeneityстатистическое распознавание образоввероятностная нейронная сетьпроверка гипотезы об однородности выборокпринцип минимума информационного рассогласования Кульбака-ЛейблераKullback-Leibler minimum information discrimination principle
Publication based on the results of:
Savchenko A., Neural Networks 2013 Vol. 46 P. 227-241
The article is devoted to pattern recognition task with the database containing small number of samples per class. By mapping of local continuous feature vectors to a discrete range, this problem is reduced to statistical classification of a set of discrete finite patterns. It is demonstrated that Bayesian decision under the assumption that probability distributions ...
Added: June 16, 2013
Savchenko A., Lecture Notes in Computer Science 2016 Vol. 9719 P. 505-512
Probabilistic neural network (PNN) is the well-known instance-based learning algorithm, which is widely used in various pattern classification and regression tasks, if rather small number of instances for each class is available. The known disadvantage of this network is its insufficient classification computational complexity. The common way to overcome this drawback is the reduction techniques ...
Added: July 6, 2016
Savchenko A., Optical Memory and Neural Networks (Information Optics) 2013 Vol. 22 No. 3 P. 184-192
The research subject is the computational complexity of the probabilistic neural network (PNN) in the pattern recognition problem for large model databases. We examined the following methods of increasing the efficiency of a neuralnetwork classifier: a parallel multithread realization, reducing the PNN to a criterion with testing of homogeneity of feature histograms of input and ...
Added: September 10, 2013
Savchenko A., Компьютерная оптика 2018 Т. 42 № 1 С. 149-158
In this paper we study the image recognition tasks, in which images are described by high dimensional feature vectors extracted with deep convolutional neural networks and principal component analysis. In particular, we focus on the problem of high computational complexity of statistical approach with non-parametric estimates of probability density implemented by the probabilistic neural network. ...
Added: April 11, 2018
Savchenko A., , in : Artificial Neural Networks in Pattern Recognition 5th INNS IAPR TC 3 GIRPR Workshop, ANNPR 2012, Trento, Italy, September 2012 Proceeding. : Berlin, Heidelberg : Springer, 2012. P. 93-103.
Since the works by Specht, the probabilistic neural networks (PNNs) have attracted researchers due to their ability to increase training speed and their equivalence to the optimal Bayesian decision of classification task. However, it is known that the PNN's conventional implementation is not optimal in statistical recognition of a set of patterns. In this article ...
Added: September 21, 2012
Savchenko A., Computer Optics 2013 Vol. 37 No. 2 P. 254-262
The usage of the probabilistic neural network with homogeneity testing is proposed in image recognition problem. This decision is shown to be optimal in Bayesian terms if the task is formulated as a statistical testing for homogeneity of query and model images' feature sets. The problem of the lack of computing efficiency with many classes ...
Added: July 1, 2013
Savchenko A., Lecture Notes in Computer Science 2014 Vol. 8641 P. 261-266
Conventional image recognition methods usually include dividing the keypoint neighborhood (for local features) or the whole object (for global features) into a grid of blocks, computing the gradient magnitude and orientation at each image sample point and uniting the orientation histograms of all blocks into a single descriptor. The query image is recognized by matching ...
Added: August 27, 2014
Savchenko A., Заводская лаборатория. Диагностика материалов 2014 Т. 80 № 3 С. 70-80
Рассмотрена задача статистического распознавания объектов со сложной структурой. Предложена математическая модель образа как последовательности выборок независимых одинаково распределенных векторов признаков. На основе этой модели и классического байесовского подхода задача распознавание сведена к проверке гипотез об однородности выборок. Представлены результаты экспериментального исследования в задачах классификации лиц и распознавании изолированных слов русской речи. ...
Added: April 8, 2014
Savchenko A., Belova N. S., / Cornell University. Series "Working papers by Cornell University". 2017.
The paper deals with the still-to-video face recognition for the small sample size problem based on computation of distances between high-dimensional deep bottleneck features. We present the novel statistical recognition method, in which the still-to-video recognition task is casted into Maximum A Posteriori estimation. In this method we maximize the joint probabilistic density of the ...
Added: August 29, 2017
Savchenko A., Компьютерная оптика 2017 Т. 41 № 3 С. 422-430
In this paper we focus on the image recognition problem in the case of small sample size based on the nearest neighbor rule and matching of high-dimensional feature vectors extracted with the deep convolutional neural network. We propose the novel recognition algorithm based on the maximum likelihood method for the joint density of dissimilarities between ...
Added: July 8, 2017
Savchenko A., Информационные системы и технологии 2014 № 2 С. 12-18
Предложен новый критерий сегментации речи, основанный на идее вероятностной нейронной сети с проверкой однородности. Экспериментально продемонстрировано, что предложенный подход позволяет на 2-7% повысить точность распознавания гласных звуков в слоге по сравнению с традиционным критерием, основанным на сопоставлении с фиксированным порогом расстояния между очередным фреймом и предыдущим однородным участком ...
Added: March 26, 2014
Savchenko A., Lecture Notes in Computer Science 2015 Vol. 9124 P. 236-245
The insufficient performance of statistical recognition of composite objects (images, speech signals) is explored in case of medium-sized database (thousands of classes). In contrast to heuristic approximate nearest-neighbor methods we propose a statistically optimal greedy algorithm. The decision is made based on the Kullback-Leibler minimum information discrimination principle. The model object to be checked at ...
Added: July 5, 2015
Savchenko A., Lecture Notes in Artificial Intelligence 2012 Vol. 7477 LNCS P. 93-103
Since the works by Specht, the probabilistic neural networks (PNNs) have attracted researchers due to their ability to increase training speed and their equivalence to the optimal Bayesian decision of classification task. However, it is known that the PNN's conventional implementation is not optimal in statistical recognition of a set of patterns. In this article ...
Added: February 26, 2013
Savchenko A., IEEE Transactions on Neural Networks and Learning Systems 2020 Vol. 31 No. 2 P. 651-660
If the training data set in image recognition task is not very large, the feature extraction with a convolutional neural network is usually applied. Here, we focus on the nonparametric classification of extracted feature vectors using the probabilistic neural network (PNN). The latter is characterized by the high runtime and memory space complexity. We propose ...
Added: November 1, 2019
Savchenko A., Belova N. S., Savchenko Lyudmila V., Optical Memory and Neural Networks (Information Optics) 2018 Vol. 27 No. 1 P. 23-31
We discuss the video classification problem with the matching of feature vectors extracted using deep convolutional neural networks from each frame. We propose the novel recognition method based on representation of each frame as a sequence of fuzzy sets of reference classes whose degrees of membership are defined based on asymptotic distribution of the Kullback–Leibler ...
Added: February 9, 2018
Savchenko A., Milov V., Optical Memory and Neural Networks (Information Optics) 2016 Vol. 25 No. 2 P. 79-87
Decision support in equipment condition monitoring systems with image processing is analyzed. Long-run accumulation of information about earlier made decisions is used to realize the adaptiveness of the proposed approach. It is shown that unlike conventional classification problems, the recognition of abnormalities uses training samples supplemented with reward estimates of earlier decisions and can be ...
Added: July 10, 2016
Savchenko A., Savchenko L. V., Lecture Notes in Artificial Intelligence 2013 Vol. 7911 P. 176-183
The definition of a phoneme as a fuzzy set of minimal speech units from the model database is proposed. On the basis of this definition and the Kullback-Leibler minimum information discrimination principle the novel phoneme recognition algorithm has been developed as an enhancement of the phonetic decoding method. The experimental results in the problems of ...
Added: June 16, 2013
Savchenko A., Pattern Recognition 2017 Vol. 61 P. 459-469
An exhaustive search of all classes in pattern recognition methods cannot be implemented in real-time, if the database contains a large number of classes. In this paper we introduce a novel probabilistic approximate nearest-neighbor (NN) method. Despite the most of known fast approximate NN algorithms, our method is not heuristic. The joint probabilistic densities (likelihoods) ...
Added: August 30, 2016
Savchenko A., Knowledge-Based Systems 2016 Vol. 91 P. 252-262
The paper is focused on an application of sequential three-way decisions and granular computing to the problem of multi-class statistical recognition of the objects, which can be represented as a sequence of independent homogeneous (regular) segments. As the segmentation algorithms usually make it possible to choose the degree of homogeneity of the features in a ...
Added: December 4, 2015
Savchenko A., Информационные системы и технологии 2015 № 4(90) С. 28-38
Рассмотрена проблема недостаточной вычислительной эффективности вероятностной нейронной сети (ВНС) в задачах распознавания образов при наличии в базе данных для каждого класса небольшого числа эталонов. На основе проекционных оценок плотности распределения с ядром Фейера и наивного предположения о независимости признаков классифицируемого объекта синтезирована новая модификация ВНС. Экспериментально показано, что предложенный классификатор оказался несколько точнее и намного ...
Added: October 8, 2015
Savchenko A., Automation and Remote Control 2016 Vol. 77 No. 3 P. 443-450
We study the recognition problem for composite objects based on a probabilistic model of a piecewise regular object with thousands of alternative classes. Using the model’s asymptotic properties, we develop a new maximal likelihood enumeration method which is optimal (in the sense of choosing the most likely reference for testing on every step) in the ...
Added: April 11, 2016
Savchenko A., Belova N. S., Expert Systems with Applications 2018 Vol. 108 P. 170-182
The paper deals with unconstrained face recognition task for the small sample size problem based on computation of distances between high-dimensional off-the-shelf features extracted by deep convolution neural network. We present the novel statistical recognition method, which maximizes the likelihood (joint probabilistic density) of the distances to all reference images from the gallery set. This ...
Added: May 17, 2018
Savchenko A., Акатьев Д. Ю., Информационные системы и технологии 2013 № 3 (77) С. 5-12
Рассмотрена проблема вариативности разговорной речи в задаче формирования фонетической базы данных. Для её решения предложено использование автоматической сегментации речи на последовательность фонем на основе когнитивной акустической модели типа фонетического кластера, определённого на множестве минимальных звуковых единиц. Разработан адаптивный алгоритм наполнения каждого кластера одноименными минимальными звуковыми единицами из непрерывного потока речи диктора. Представлены результаты экспериментального исследования ...
Added: May 7, 2013
Savchenko A., Switzerland : Springer, 2016
A unified methodology for categorizing various complex objects is presented in this book. Through probability theory, novel asymptotically minimax criteria suitable for practical applications in imaging and data analysis are examined including the special cases such as the Jensen-Shannon divergence and the probabilistic neural network. An optimal approximate nearest neighbor search algorithm, which allows faster ...
Added: April 12, 2016