Gabor filter approximation through a Neural Network applied to fingerprint images

2016 ◽  
Author(s):  
Airam Carlos Pais Barreto Marques ◽  
Antonio Carlos Gay Thomé
Author(s):  
D. Lebedev ◽  
A. Abzhalilova

Currently, biometric methods of personality are becoming more and more relevant recognition technology. The advantage of biometric identification systems, in comparison with traditional approaches, lies in the fact that not an external object belonging to a person is identified, but the person himself. The most widespread technology of personal identification by fingerprints, which is based on the uniqueness for each person of the pattern of papillary patterns. In recent years, many algorithms and models have appeared to improve the accuracy of the recognition system. The modern algorithms (methods) for the classification of fingerprints are analyzed. Algorithms for the classification of fingerprint images by the types of fingerprints based on the Gabor filter, wavelet - Haar, Daubechies transforms and multilayer neural network are proposed. Numerical and results of the proposed experiments of algorithms are carried out. It is shown that the use of an algorithm based on the combined application of the Gabor filter, a five-level wavelet-Daubechies transform and a multilayer neural network makes it possible to effectively classify fingerprints.


2021 ◽  
Vol 7 (4) ◽  
pp. 117
Author(s):  
Linling Fang ◽  
Yingle Fan

<p>A biomimetic vision computing model based on multi-level feature channel optimization coding is proposed and applied to image contour detection, combining the end-to-end detection method of full convolutional neural network and the traditional contour detection method based on biological vision mechanism. Considering the effectiveness of the Gabor filter in perceiving the scale and direction of the image target, the Gabor filter is introduced to simulate the multi-level feature response on the visual path. The optimal scale and direction of the Gabor filter are obtained based on the similarity index, and they are used as the frequency separation parameter of the NSCT transform. The contour sub-image obtained by the NSCT transform is combined with the original image for feature enhancement and fusion to realize the primary contour response. The low-dimensional and low-redundancy primary contour response is used as the input sample of the network model to relieve network pressure and reduce computational complexity. A fully improved convolutional neural network model is constructed for multi-scale training, through feature encoder to feature decoder, to achieve end-to-end pixel prediction, and obtain a complete and continuous detection image of the subject contour. Using the BSDS500 atlas as the experimental sample, the average accuracy index is 0.85, which runs on the device CPU at a detection rate of 20+ FPS to achieve a good balance between training efficiency and detection effect.</p>


2011 ◽  
Vol 20 (03) ◽  
pp. 489-509 ◽  
Author(s):  
BEHZAD HELLI ◽  
MOHSEN EBRAHIMI MOGHADDAM

The behavioral-biometrics methods of writer identification and verification have been considered as a research topic for many years. However, many writer identification and verification methods have been designed based on English handwriting properties, but because of many differences between English and Persian handwriting and the challenges facing Persian handwriting analysis, designing such methods has many interests in Persian yet. In this paper, we have presented a fully text-independent and texture based method for identifying writers of Persian handwritten documents. As a result of special properties of Persian handwriting, a modified version of Gabor filter that is called Extended Gabor (XGabor) filter has been used to extract the features. An MLP (Multi Layer Perceptron (Node)) neural network and a K-NN classifier have been employed to classify the extracted features. In the evaluation phase, an exhaustive database of Persian handwritten documents was prepared and the method applied on. The experimental results showed that the accuracy of proposed method is about 97% and it is competitive with others. We believe that the proposed method may be extended to identify writers in other languages by adjusting some parameters.


SinkrOn ◽  
2021 ◽  
Vol 5 (2) ◽  
pp. 314-324
Author(s):  
Mawaddah Harahap ◽  
Valencia Angelina ◽  
Fenny Juliani ◽  
Celvin Celvin ◽  
Oscar Evander

Grapes are one type of fruit that is usually used to make grape juice, jelly, grapes, grape seed oil and raisins, or to be eaten directly. So far, checking for disease in grapes is still done manually, by checking the leaves of the grapes by experts. This method certainly takes a long time considering the extent of the vineyards that must be evaluated. To solve this problem, it is necessary to apply a method of detecting grape disease, so that it can help the common people to detect grape disease. This research will use the Dual-Channel Convolutional Neural Network method. The process of detecting grape disease using the DCCNN method will begin with the extraction of the leaves from the input image using the Gabor Filter method. After that, the Segmentation Based Fractal Co-Occurrence Texture Analysis method will be used to extract the features, color, and texture of the extracted leaves. The result is the number of datasets will affect the accuracy of the results of disease identification using the DCCNN method. However, more datasets will cause the execution process to take longer. Changes in the angle and frequency values in the Gabor method at the time of testing will reduce the accuracy of the test results. The conclusion of this study are the DCCNN method can be used to detect the type of leaf disease in grapes and the number of datasets will affect the accuracy of the results of disease identification using the DCCNN method.


Sign in / Sign up

Export Citation Format

Share Document