Optimal Gabor Filter Bank Research Articles

Real-time product quality control system using optimized Gabor filter bank

Machine vision systems provide significant advantages when compared to conventional methods. Inspired from this idea, a visual inspected system that is independent of dust and dirt is presented for glass production systems. The method consists of a camera, a conveyor system, and an image segmentation method. Architecture of proposed system is as follows: a specific area of the conveyor is isolated from the outside. When glass enters this area, defect inspection process begins. In the inspection process, damaged and undamaged regions on glass surface are segmented. Gabor filter is very effective to detect orientation and thickness of these defects. But, Gabor filter bank should be created using appropriate Gabor coefficients for real-time applications. Otherwise, the processing time will be too long or fail results will be obtained. For this purpose, a new Gabor filter bank is created using gray wolf optimizer. In the hardware section, light beams are injected into the glass and the movements of these beams are observed to increase the perceptibility of the damage. Beam distribution is homogenous in the undamaged regions, but homogeneity is disturbed in defected areas. To avoid irregular glare on the glass surface, external lights are blocked and an artificial light source is used. Artificial light beams are injected into perpendicularly in the glass. So, homogeneous illumination in the glass can be occurred. Finally, optimized Gabor filter bank is applied to glass images. Proposed system detects all defects on the glass surface in the experiments. Size of the smallest defect is 0.4 mm. Defect detection performance of proposed system is nearly 100%. If it is evaluated in terms of shape and size, accuracy rate is 98.1%.

Optimized Gabor features for mass classification in mammography

Gabor filter bank has been successfully used for false positive reduction problem and the discrimination of benign and malignant masses in breast cancer detection. However, a generic Gabor filter bank is not adapted to multi-orientation and multi-scale texture micro-patterns present in the regions of interest (ROIs) of mammograms. There are two main optimization concerns: how many filters should be in a Gabor filter band and what should be their parameters. Addressing these issues, this work focuses on finding optimizing Gabor filter banks based on an incremental clustering algorithm and Particle Swarm Optimization (PSO). We employ an SVM with Gaussian kernel as a fitness function for PSO. The effect of optimized Gabor filter bank was evaluated on 1024 ROIs extracted from a Digital Database for Screening Mammography (DDSM) using four performance measures (i.e., accuracy, area under ROC curve, sensitivity and specificity) for the above mentioned mass classification problems. The results show that the proposed method enhances the performance and reduces the computational cost. Moreover, the Wilcoxon signed rank test over the significance level of 0.05 reveals that the performance difference between the optimized Gabor filter bank and non-optimized Gabor filter bank is statistically significant.

Minimal Gabor Filter Bank for Face Features Localization

We propose an optimized Gabor filter bank for facial features localization. Our analysis shows that the responses to filters with selected parameters offer significantly higher values of certain statistics. The highest values of standard deviation and energy confirm their relevance as descriptors for classification. Compared to the banks with different numbers of filters, the proposed bank of 8 filters offers the best classification sensitivity using MLP.

Cloth Defect Classification Method Based on SVM

To solve the problem of low classification accuracy in cloth defect detection, an approach combining Gabor filter and SVM (Support Vector Machine) is proposed. Firstly, the optimal Gabor filter bank based on Half-peak tangent method is adopted to filter the original cloth image, and the defect is segmented out. Then, the SVM classifier is trained using the extracted defect features. In the stage of training, the best penalty factor and the Kernel parameter in SVM are searched by the Genetic Algorithm. At last, the performance of the classifier is tested. Experiment results show the proposed approach has strong detection and classification capacities to the monochrome texture defect.

Fabric Defect Detection Method Based on Optimal Gabor Filter Bank

A method of fabric defect detection online based on Gabor filter bank is proposed. First, optimal Gabor filter bank is designed to match texture features of normal fabric image and built adopting the half-peak tangent method. The filter bank’s parameters are obtained by using texture features. Then, fabric defect image is treated using this filter bank, and filtered sub-images are obtained which characterize the fabric defect in different orientations and frequencies. Finally, these sub-images are fused to reconstruct a binary defect image. And the binary defect image is output to finish the detection. Experiment results show that the Gabor filter bank based on texture features can reduce the number of filters effectively and detect most of the fabric defects.

Analysis of variance of Gabor filter banks parameters for optimal face recognition

Gabor filter banks constitute a very robust tool to extract discriminant information from a visual scene. After the now “classical” bank with 5 frequencies and 8 orientations proposed by Lades et al. and Wiskott et al., many other parametrizations of a Gabor filter bank have appeared. In order to find the optimal parametrization for a face recognition experiment, we have performed a 6-way analysis of variance of Gabor parameters using FERET, FRAV2D, FRAV3D, FRGC and XM2VTS face databases, including frontal and turned poses, facial expressions, occlusions and changes of illumination. Considering independent criteria to find the optimal Gabor filter bank, the bank with the highest recognition rate was found to have 6 frequencies and narrower Gaussian widths in the space domain. These results were obtained with Mahalanobis distance for a k-NN classifier, with analytical and holistic Gabor feature vectors. Moreover about 20% of the banks studied here obtained in average a better performance than the classical bank. For most of the databases considered, the highest recognition rates have been achieved with analytical representations (frontal images, images with turns or occlusions), with a holistic preponderance for images with gestures or changes of illumination. The inferiority found for holistic Gabor representations versus their analytical counterparts can be explained for the intrinsic redundancy and the size of the feature vectors of this kind of representation.

Iris recognition using Gabor filters optimized by the particle swarm algorithm

An efficient feature extraction algorithm based on optimized Gabor filters and a relative variation analysis approach is proposed for iris recognition. The Gabor filters are optimized by using the particle swarm algorithm to adjust the parameters. Moreover, a sequential scheme is developed to determine the number of filters in the optimal Gabor filter bank. In the preprocessing step, the lower part of the iris image is unwrapped and normalized to a rectangular block that is then decomposed by the optimal Gabor filters. After that, a simple encoding method is adopted to generate a compact iris code. Experimental results show that with a smaller iris code size, the proposed method can produce comparable performance to that of the existing iris recognition systems.

2D Gaborface representation method for face recognition with ensemble and multichannel model

This paper proposes a scheme that is based on linear correlation criterion to select optimized Gabor filter bank. In addition, by using 2D Gaborface matrices rather than transformed 1D feature vectors, a novel Gaborface-based 2DPCA and (2D) 2PCA classification method is introduced. Two kinds of strategies to use the bank of Gaborfaces are proposed: ensemble Gaborface representation (EGFR) and multichannel Gaborface representation (MGFR). The feasibility of our method is proved with the experimental results on the ORL, Yale and FERET databases. In particular, the MGFR-based (2D) 2PCA method achieves 100% recognition accuracy for ORL database, and 98.89% accuracy for Yale database with five training samples per class, and 99.5% accuracy for FERET database.