Gabor Filters Research Articles

Atypical architectural distortion detection in digital breast tomosynthesis: a computer-aided detection model with adaptive receptive field

Objective. In digital breast tomosynthesis (DBT), architectural distortion (AD) is a breast lesion that is difficult to detect. Compared with typical ADs, which have radial patterns, identifying a typical ADs is more difficult. Most existing computer-aided detection (CADe) models focus on the detection of typical ADs. This study focuses on atypical ADs and develops a deep learning-based CADe model with an adaptive receptive field in DBT. Approach. Our proposed model uses a Gabor filter and convergence measure to depict the distribution of fibroglandular tissues in DBT slices. Subsequently, two-dimensional (2D) detection is implemented using a deformable-convolution-based deep learning framework, in which an adaptive receptive field is introduced to extract global features in slices. Finally, 2D candidates are aggregated to form the three-dimensional AD detection results. The model is trained on 99 positive cases with ADs and evaluated on 120 AD-positive cases and 100 AD-negative cases. Main results. A convergence-measure-based model and deep-learning model without an adaptive receptive field are reproduced as controls. Their mean true positive fractions (MTPF) ranging from 0.05 to 4 false positives per volume are 0.3846 ± 0.0352 and 0.6501 ± 0.0380, respectively. Our proposed model achieves an MTPF of 0.7148 ± 0.0322, which is a significant improvement (p < 0.05) compared with the other two methods. In particular, our model detects more atypical ADs, primarily contributing to the performance improvement. Significance. The adaptive receptive field helps the model improve the atypical AD detection performance. It can help radiologists identify more ADs in breast cancer screening.

An Intelligent Hybrid Approach for Brain Tumor Detection

Brain tumours are quickly increasing in prevalence all over the world. It causes the deaths of thousands of individuals annually. Misdiagnosis of brain tumours often results in unnecessary treatment, further lowering the survival rate of the affected individuals. Prompt medical diagnosis is crucial to improve the prognosis for patients with brain tumours. Positive advancements in deep and machine learning domains have been made due to repeated achievements in supporting medical practitioners in making correct diagnoses utilizing computer-aided diagnostic tools. Deep convolutional layers are superior to conventional methods at extracting unique characteristics from target regions. In this research, initially, Gabor filter and ResNet50 were applied to accurately extract the important features of brain tumours from the MRI images dataset. Firstly, the extracted features of Gabor and ResNet50 were classified individually through SVM, and secondly, the features from both these techniques were combined and then classified through SVM. The Kaggle MRI dataset for a brain tumour was utilized in this research. It includes 7,023 Images and four classes Glioma, Meningioma, No-Tumor, and Pituitary. The results from every system were outstanding, but the best results were shown by the combined features of Gabor and ResNet50, an advanced hybrid approach with 95.73% accuracy, 95.90% precision, and 95.72% f1 score.&#x0D;

Research on Vessel Segmentation and Extraction Method Based on Improved 2D Gabor Features

This study proposed a vessel segmentation method based on Gabor features. According to the eigenvector of Hessian matrix of each pixel in the image, the vessel direction of each point was obtained to set the direction angle of Gabor filter, and the Gabor features of different vessel width at each point were extracted to establish the 6D vectors of each point. By reducing the dimension of the 6D vector, the 2D vector of each point was obtained and fused with the original image G channel. U-Net neural network was used to classify the fused image to segment vessels. The experimental results of this method in DRIVE dataset showed that this method had a good effect on the detection of small vessels and vessels at the intersection.

Multi-source perceptual blind compensation inspection method for substation based on equipment’s visual blind area identification and saliency detection

Abstract In order to expand the detection range and ensure the operation stability, the substation multi-source perception blind compensation detection method based on equipment visual blind area recognition and significance detection is studied. Acoustic sensors are used to collect acoustic signals from visual blind areas of equipment. The characteristics of noise signal are identified by wavelet analysis and noise reduction. The supercomplex Fourier transform model was used to extract the important region in the device image, and the texture features of the region were detected by Gabor filter. The blind compensation detection feature vector is formed by integrating two multi-source sensing features. The detection model of support vector machine is input to complete the blind compensation detection of the substation. The experimental results show that the proposed method is effective for the sound signal feature recognition in the visual blind area and the texture feature detection in the significant area of the device image. The different operating states of each equipment detected by the multi-source sensing feature vector are more accurate, which can realize the purpose of the multi-source sensing blind compensation check of the substation and ensure the safe and stable operation of the substation.

Semantic segmentation using GSAUNet

Gabor filters well support computer vision tasks due to its steerable properties, that are useful for handling image transformation such as image rotation and image scaling. In this paper, the Gabor-based UNet (GSAUNet) model is proposed where the Gabor filters replace the few default filters in the UNet. Besides, it has adopted a squeeze-and-excitation module to increase channel interdependencies and attention mechanism to focus on an important part of the network. The performance of GSAUNet is assessed on the CamVid and Cityscape datasets that attain improvement for the semantic segmentation tasks.

Iris recognition based on 2D Gabor filter

&lt;span lang="EN-US"&gt;Iris recognition is a type of biometrics technology that is based on physiological features of the human body. The objective of this research is to recognize and identify iris among many irises that are stored in a visual database. This study employed a left and right iris biometric framework for inclusion decision processing by combining image processing and artificial bee colony. The proposed approach was evaluated on a visual database of 280 colored iris pictures. The database was then divided into 28 clusters. Images were preprocessed and texture features were extracted based Gabor filters to capture both local and global details within an iris. The technique begins by comparing the attributes of the online-obtained iris picture with those of the visual database. This technique either generates a reject or approve message. The consequences of the intended work reflect the output’s accuracy and integrity. This is due to the careful selection of attributes, as well as the deployment of an artificial bee colony and data clustering, which decreased complexity and eventually increased identification rate to 100%. We demonstrate that the proposed method achieves state-of-the-art performance and that our recommended procedures outperform existing iris recognition systems.&lt;/span&gt;

Demarcating Z-line and Gastric Folds Boundary Based on the Segmentation of the Lower Esophageal Sphincter Images.

The endoscopic diagnosis of pathological changes in the gastroesophageal junction including esophagitis and Barrett's mucosa is based on the visual detection of two boundaries: mucosal color change between esophagus and stomach, and top endpoint of gastric folds. The presence and pattern of mucosal breaks in the gastroesophageal mucosal junction (Z line) classify esophagitis in patients and the distance between the two boundaries points to the possible columnar lined epithelium. Since visual detection may suffer from intra- and interobserver variability, our objective was to define the boundaries automatically based on image processing algorithms, which may enable us to measure the detentions of changes in future studies. To demarcate the Z-line, first the artifacts of endoscopy images are eliminated. In the second step, using SUSAN edge detector, Mahalanobis distance criteria, and Gabor filter bank, an initial contour is estimated for the Z-line. Using region-based active contours, this initial contour converges to the Z-line. Finally, by applying morphological operators and Gabor Filter Bank to the region inside of the Z-line, gastric folds are segmented. To evaluate the results, a database consisting of 50 images and their ground truths were collected. The average dice coefficient and mean square error of Z-line segmentation were 0.93 and 3.3, respectively. Furthermore, the average boundary distance criteria are 12.3 pixels. In addition, two other criteria that compare the segmentation of folds with several ground truths, i.e., Sweet-Spot Coverage and Jaccard Index for Golden Standard, are 0.90 and 0.84, respectively. Considering the results, automatic segmentation of Z-line and gastric folds are matched to the ground truths with appropriate accuracy.

Fruit Type Recognition Using Hybrid Method with Principal Component Analysis (PCA)

This research concentrated on fruit image recognition. Fruit recognition in this study can be used to estimate the number of fruits that exist. Data testing is used to classify a fruit image that has been trained to recognize a variety of labels (fruit types). Until the classification process, several processes and methods are used in this research, one of which is the Gaussian filter to improve the quality of fruit image recognition. In addition, the Gabor filter is used in the feature extraction process, and the PCA technique is used in feature selection to select the best features. To classify the chosen feature, deep learning and the k-nearest neighbor (k-NN) method will be used. Furthermore, the processes used improved the accuracy of 92%, a root mean squared error (RMSE) of 0.323, a mean squared error (MSE) of 0.6278.

Sputum smears quality inspection using an ensemble feature extraction approach.

The diagnosis of tuberculosis (TB) is extremely important. Sputum smear microscopy is thought to be the best method available in terms of accessibility and ease of use in resource-constrained countries. In this paper, research was conducted to evaluate the effectiveness of tuberculosis diagnosis by examining, among other things, the underlying causes of sputum smear quality for Ethiopian states such as Tigray, Amahira, and Oromia. However, because it is done manually, it has its limitations. This study proposes a model for sputum smear quality inspection using an ensemble feature extraction approach. The dataset used was recorded and labeled by experts in a regional lab in Bahir Dar, near Felege Hiwot Hospital after being collected from Gabi Hospital, Felege Hiwot Hospital, Adit Clinic and Gondar Hospital, as well as Kidanemihret Clinic in Gondar. We used a controlled environment to reduce environmental influences and eliminate variation. All the data was collected using a smartphone (the standard 15) with a jpg file extension and a pixel resolution of 1,728 × 3,840. Prior to feature extraction, bicubic resizing, and ROI extraction using thresholding was performed. In addition, sequential Gaussian and Gabor filters were used for noise reduction, augmentation, and CLAHE was used for enhancement. For feature extraction, GLCM from the gray label and CNN from the color image were both chosen. Ultimately, when CNN, SVM, and KNN classifiers were used to test both CNN and GLCM features, KNN outperformed them all with scores of 87, 93, and 94% for GLCM, CNN, and a hybrid of CNN and GLCM, respectively. CNN with GLCM outperformed other methods by 0.7 and 0.1% for GLCM and CNN feature extractors using the same classifier, respectively. In addition, the KNN classifier with the combination of CNN and GLCM as feature extractors performed better than existing methods by 1.48%.

Deteksi Kecacatan Permukaan Rel Menggunakan Metode Deep Learning Neural Network

Rail is a construction in one unit made of steel, concrete, and other construction materials above or below the ground depending on the direction and area. The condition of the rail surface must not have any defects so that train travel is safe and comfortable. This research method discusses the research design. There are four processes in this research design, namely the data acquisition process, the RGB image color conversion process to HSI, the filtering process using the gabor filter, and the classification process using deep learning neural networks. The purpose of this study is to build a system to visually detect defects in the surface of the railroad tracks, namely image processing techniques. This activity was carried out at the Madiun Indonesian Railways Polytechnic Station Laboratory. Based on the research that has been done, it can be concluded that an image with a size of 32x32 pixels produces the highest accuracy value at epoch 90 using the gabor filter image type. The more epochs used, the better the results will be and the better the model can be made. The accuracy results obtained were 0.8041 or 80.41% for training accuracy and testing accuracy of 0.79 or 79%

Explaining the physics of transfer learning in data-driven turbulence modeling.

Transfer learning (TL), which enables neural networks (NNs) to generalize out-of-distribution via targeted re-training, is becoming a powerful tool in scientific machine learning (ML) applications such as weather/climate prediction and turbulence modeling. Effective TL requires knowing (1) how to re-train NNs? and (2) what physics are learned during TL? Here, we present novel analyses and a framework addressing (1)-(2) for a broad range of multi-scale, nonlinear, dynamical systems. Our approach combines spectral (e.g. Fourier) analyses of such systems with spectral analyses of convolutional NNs, revealing physical connections between the systems and what the NN learns (a combination of low-, high-, band-pass filters and Gabor filters). Integrating these analyses, we introduce a general framework that identifies the best re-training procedure for a given problem based on physics and NN theory. As test case, we explain the physics of TL in subgrid-scale modeling of several setups of 2D turbulence. Furthermore, these analyses show that in these cases, the shallowest convolution layers are the best to re-train, which is consistent with our physics-guided framework but is against the common wisdom guiding TL in the ML literature. Our work provides a new avenue for optimal and explainable TL, and a step toward fully explainable NNs, for wide-ranging applications in science and engineering, such as climate change modeling.

Detection and classification of lung cancer computed tomography images using a novel improved deep belief network with Gabor filters

The computer-aided diagnosis (CAD) method plays a considerable role in the automated recognition of medical images, considering the increasing numbers of lung cancer patients. Pulmonary nodules are a type of lung irregularity that can be detected early in the course of a patient's life using CAD-based pulmonary patient diagnosis. For several classification methods of lung CT images, the existing method has a lower performance in early cancer diagnosis and a longer processing time. To improve the performance of the classification method in terms of processing time and accuracy, we propose a novel method to classify lung CT images. In the proposed method of lung cancer classification, we use Gabor filters with an enhanced Deep Belief Network (E-DBN) with multiple classification methods. In this E-DBN, we use two cascaded RBMs, Gaussian-Bernoulli (GB) and Bernoulli–Bernoulli (BB) RBMs. A support vector machine (SVM) gives the best performance parameters of all applied methods. Our proposed model uses a support vector machine with an E-DBN to classify lung CT images and improve performance parameters like accuracy, sensitivity, specificity, F-1 score, false positive rate (FPR), false negative rate (FNR), and ROC curve. The proposed approaches are tested and evaluated using three publicly available datasets: the LIDC-IDRI and LUNA-16 datasets.

(Retracted) Simulating human visual perception behaviors for video quality assessment

Multimedia video quality evaluation system plays an important role in modern Internet applications, such as WebRTC-based video conference. However, video quality may be affected during transmission because of noise or network congestion. Constructing an end-to-end intelligent video transmission monitoring system is of great significance. We design a system based on no-reference quality assessment, which can be used to evaluate the video-based object quality evaluation. More specifically, we first leverage Gabor filter for video segmentation. Then we conduct image smoothness estimation to measure the correlation between two adjacent frames. Subsequently, we consider two states of video distortion: compression and video transmission. We model these two states, respectively. We compare our VQA algorithm with a set of well-known NR-IQA. The reported VQA accuracies have clearly shown the superiority of our proposed video quality predictor.

Multi-scale lung tissue classification for interstitial lung diseases using learned Gabor filters

Multi-scale lung tissue classification for interstitial lung diseases using learned Gabor filters

Lumbar intervertebral disc segmentation for computer modeling and simulation

Background and objectiveThe present work had as its main objective the development of a method for localizing and automatically segmenting lumbar intervertebral discs (IVD) in 3D from magnetic resonance imaging (MRI), with the goal of supporting the generation of finite element (FE) models from actual lumbar spine anatomy, by providing accurate and personalized information on the shape of the patient's IVD. The extension of the method to allow performing separate segmentations of the IVD's two main structures – annulus fibrosus (AF) and nucleus pulposus (NP) – as well as automatically detecting degenerated IVD where this distinction is no longer possible was also an objective of the work. MethodsThe method presented here evolves from 2D segmentations in the sagittal profile using Gabor filters towards 3D segmentations. It works by detecting the spine curves and intensity regions corresponding to IVD. As so, the 2D method from Zhu et al. (2013) was partially implemented, modified and adapted to 3D use, and then tested with eight spines from two separated online datasets. The 3D adaptation was achieved by using vertebral body segmentation masks to approximate the shape of the vertebrae and to adjust the spine curves accordingly. ResultsThe method showed average values of 85%, 83% and 96% for the Dice coefficient, sensitivity and specificity, respectively. The method correctly identified 65 of 68 (96%) IVD as either healthy or degenerated. The method's Dice coefficient is within the range of existing 3D IVD segmentation methods in the literature (81–92%). The method took on average 6–7 s to perform a full 3D segmentation, which is well within the range of the existing methods (2 s – 19 min). ConclusionsThe developed method can be used to generate accurate 3D models of the IVD based on MRI, with AF/NP distinction and detection of marked degeneration by comparing each IVD with the remaining spine levels. Further work shall improve the method towards distinguishing between specific levels of degeneration for clinically oriented FE modeling.

Unsupervised Learning for the Segmentation of Small Crystalline Particles at the Atomic Level

AbstractElectron backscattering diffraction provides the analysis of crystalline phases at large scales (microns) while precession electron diffraction may be used to get 4D‐STEM data to elucidate structure at nanometric resolution. Both are limited by the probe size and also exhibit some difficulties for the generation of large datasets, given the inherent complexity of image acquisition. The latter appoints the application of advanced machine learning techniques, such as deep learning adapted for several tasks, including pattern matching, image segmentation, etc. This research aims to show how Gabor filters provide an appropriate feature extraction technique for electron microscopy images that could prevent the need of large volumes of data to train deep learning models. The work presented herein combines an algorithm based on Gabor filters for feature extraction and an unsupervised learning method to perform particle segmentation of polyhedral metallic nanoparticles and crystal orientation mapping at atomic scale. Experimental results have shown that Gabor filters are convenient for electron microscopy images analysis, that even a nonsupervised learning algorithm can provide remarkable results in crystal segmentation of individual nanoparticles. This approach enables its application to dynamic analysis of particle transformation recorded with aberration‐corrected microscopy, offering new possibilities of analysis at nanometric scale.

Fingerprint image enhancement using multiple filters.

Biometrics is the measurement of an individual's distinctive physical and behavioral characteristics. In comparison to traditional token-based or knowledge-based forms of identification, biometrics such as fingerprints, are more reliable. Fingerprint images recorded digitally can be affected by scanner noise, incorrect finger pressure, condition of the finger's skin (wet, dry, or abraded), or physical material it is scanned from. Image enhancement algorithms applied to fingerprint images remove noise elements while retaining relevant structures (ridges, valleys) and help in the detection of fingerprint features (minutiae). Amongst the most common image enhancement filters is the Gabor filter, however, given their restricted maximum bandwidth as well as limited range of spectral information, it falls short. We put forward a novel method of fingerprint image enhancement using a combination of a diffusion-coherence filter and a 2D log-Gabor filter. The log-Gabor overcomes the limitations of the Gabor filter while Coherence Diffusion mitigates noise elements within fingerprint images. Implementation is done on the FVC image database and assessed via visual comparison with coherence diffusion used disjointedly and with the Gabor filter.

Retinal blood vessel segmentation using root Guided decision tree assisted enhanced Fuzzy C-mean clustering for disease identification

A number of groundbreaking computer-aided techniques have recently been established for the segmentation of retinal vessels, with clinical applications predominating. Early detection of these abnormalities is critical in preventing individuals from becoming blind. In recent days, automatic vessel segmentation is necessary, but it is a difficult task due to the complex structure and variation in vessel size and shape of retinal vessels. Therefore, in this work, a novel retinal Blood Vessel Segmentation (BVS) has been proposed that utilizes the potential of the Enhanced Fuzzy C-mean Clustering (EFCM) and Root Guided Decision Tree (RGTC). Initially, the input images are pre-processed to enhance the color channel contrast and the Gabor filter is utilized to enhance the small vessels. In the segmentation phase, a 13-dimensional feature vector is constructed, which is then reduced into an optimum feature set by utilizing Principal Component Analysis (PCA). The segmentation phase is utilized to further process the new feature vector. The RGDT is used to undertake reanalysis in non-vessel clustersin order to recover the incorrectly grouped vessel pixels. Finally, the accurate segmented image is obtained through post-processing morphological operations. Simulation has been performed using Python with OpenCV and an accuracy of 97.69% and 97.51% has been achieved using in STARE dataset and DRIVE dataset respectively. Experimental results indicate that the proposed algorithm delivers satisfactory results. As a result, this method is useful as an automated retinal BVS system to assist ophthalmologists.

Hyperspectral Image Classification Based on the Gabor Feature with Correlation Information

Gabor filter is widely used to extract spatial texture features of hyperspectral images (HSI) for HSI classification; however, a single Gabor filter cannot obtain the complete image features. In the paper, we propose an HSI classification method that combines the Gabor filter (GF) and domain-transformation standard convolution (DTNC) filter. First, we use the Gabor filter to extract spatial texture features from the first two principal components of the dimensionality-reduction HSI with PCA. Second, we use the DTNC filter to extract spatial correlation features from HSI in all bands. Finally, the Large Margin Distribution Machine (LDM) uses the linear fusion of the two kinds of spatial features to classify HSI. The experimental results show that the classification accuracy of Indian Pines, Pavia University, and Kennedy Space Center data sets is 96.64, 98.23, and 98.95% with only 4, 3, and 6% training samples, respectively; and these accuracies are 2–20% higher than the other tested methods. Compared with the hyperspectral information based on SVM, EPF, IFRF, PCA-EPFs, LDM-FL, and GFDN method, the proposed method, GFDTNCLDM, significantly improves the accuracy of HSI classification.

Fingerprint classification combined with Gabor filter and convolutional neural network

A fingerprint is an impression left by the friction ridges of a human finger. A fingerprint classification system groups fingerprint according to their characteristics and therefore helps to match a fingerprint against an extensive database of fingerprints. The Henry classification system is widely used among fingerprint classification systems. Some researchers have used traditional machine learning or deep learning for fingerprint classification. Nevertheless, traditional algorithms cannot extract the depth features of the fingerprint, and most deep learning algorithms lack fingerprint image enhancement. So, this paper combined the Gabor Filter and Convolutional Neural Network to extract fingerprint features. The model has two channels, one is a Deep Convolutional Neural Network (DCNN), and the other is a Shallow Convolutional Neural Network (SCNN). The DCNN consists of a neural network with eight layers, which can extract deep features of the fingerprint. The SCNN consists of Gabor Filter and a neural network with two layers that can extract features from clear fingerprint images. This paper uses NIST Special Database 4 for experiments. Experimental results show that the model proposed in this paper has achieved 91.4% accuracy. Compared with other algorithms, this model has higher accuracy than others. It shows that combined with the Gabor Filter and Convolutional Neural Network can better extract the ridge features of fingerprint images.