Showing 4 results for Pattern Recognition
, ,
Volume 10, Issue 2 (3-2014)
Abstract
In this paper, two kinds of chaotic neural networks are proposed to evaluate the efficiency of chaotic dynamics in robust pattern recognition. The First model is designed based on natural selection theory. In this model, attractor recurrent neural network, intelligently, guides the evaluation of chaotic nodes in order to obtain the best solution. In the second model, a different structure of chaotic neural network is presented which includes chaotic neurons in the hidden layer. The behavior of these neurons can be controlled by changing the parameters of chaotic neurons. Furthermore, both models are supposed to recognize the noisy patterns even those with high levels of additional noise (up to 60%). Using the first proposed model, the accuracy of recognition was improved by 37.16%, 29.15% and 8.5% comparing to feedforward neural network, chaotic neural network based on chaotic nodes - NDRAM, and ARNN respectively. The second model increased the accuracy of recognition by an average of 13.91%, and 5.41% in comparison to ARNN and first model. In addition, it has been observed that the second model had a better performance, even in point attractor mode, than ARNN which acts in non chaotic mode.
Gholam Ali Montazer , Mohammad Shayestehfar,
Volume 12, Issue 1 (6-2015)
Abstract
License plate recognition is one of the most important applications used in intelligent transportation systems. Difficulty of correct detection and identification of the car plates in different environment conditions makes researchers try new approaches to better solve the problem. License plate recognition problem is divided into three sub problems: "Plate Location", "Character Segmentation", and "Character Identification". In this paper we have tried to improve location and identification of Iranian license plate with fuzzy rules. License locating has been done with edge detection, morphological operations and using fuzzy rules and characters have been identified by fuzzy support vector machine. By applying the algorithm on 50 images, 90% of plates were located and 94% of characters were identified successfully. This shows superiority of our algorithm over non-fuzzy approaches.
Mohsen Farhang, Hosein Bahramgiri, Hamid Dehghani,
Volume 13, Issue 2 (9-2016)
Abstract
In this paper a feature-based modulation classification algorithm is developed for discriminating PSK signals. The candidate modulation types are assumed to be QPSK, OQPSK, π/4 DQOSK and 8PSK. The proposed method applies an 8PSK baseband demodulator in order to extract required features from observed symbols. The received signal with unknown modulation type is demodulated by an 8PSK demodulator whose output is considered as a finite state machine with different states and transitions for each candidate modulation. Estimated probabilities of particular transitions constitute the discriminating features. The obtained features are given to a Bayesian classifier which decides on the modulation type of the received signal. The probability of correct classification is computed with different number of observed symbols and SNR conditions by carrying out several simulations. The results show that the proposed method offers more accurate classification compared to previous methods for classifying variants of QPSK.
Mr Amir Soltany Mahboob, Mr Seyed Hamid Zahiri Mamaghani,
Volume 16, Issue 4 (3-2020)
Abstract
ANFIS systems have been much considered due to their acceptable performance in terms of creation of fuzzy classifier and training. One main challenge in designing an ANFIS system is to achieve an efficient method with high accuracy and appropriate interpreting capability. Undoubtedly, type and location of membership functions and the way an ANFIS network is trained are of considerable effect on its performance. Up to present time, related researches have just found type and location of membership functions, and or suggested methods to train these networks. Main reason for lack of simultaneous determination of type and location of membership functions and training an ANFIS network is the length of standard versions of Heuristic methods being fixed. In this paper, a new version of optimization method of inclined planes will be introduced, primarily; while search factors could be variable. Then, achieved capability will be used for specifying type and location of membership functions and simultaneous training of a classifier based on adaptive neuro-fuzzy inference system (ANFIS). The proposed method on five benchmark datasets iris, Breast Cancer, Bupa Liver, Wine and Pima from the UCI database has been tested, which has different number of reference classes, different length of attribute vectors with appropriate complexity. Initially, the accuracy of the test dataset for each of the selected datasets was compared using the standard 10 folded cross validation method using the standardized version of the standard length.Then the same experiments were repeated by the proposed method and the results of applying the proposed method on the five aforementioned datasets were compared with the results of the heuristic methods with the standard length version. The comparative results show that the optimal and intelligent design of ANFIS classifier by variable length heuristics on five well-known datasets yields good and satisfactory results and in each of the five problems it has provided better answers than other design methods in the ANFIS classification system.
