LBP Features Research Articles

Developing an extreme learning machine based approach to weed segmentation in pastures

Effective weed management in pastures is critical for maintaining the productivity of grazing land. Autonomous ground vehicles (AGVs) are increasingly being considered for weed localization and treatment in agricultural land. Weeds, however, can be difficult to distinguish from background plants, due to similarities in colour, shape and texture. While deep learning approaches can be used to solve the localization issue, they are computationally expensive, and require a large volume of training images in order to combat overfitting. In this paper we present a novel Extreme Learning Machine based network for segmenting weeds from the background pasture. The proposed method utilizes a combination of LBP, HOG and colour features, and is tested on four small datasets, achieving a high mean Intersection over Union of 87.1, 79.5, 81.6 and 87.6 for Bathurst burr, horehound, thistle and serrated tussock respectively.

Brain Tumor Detection Based on A Combination of GLCM and LBP Features with PCA and IG

Brain Tumor Detection Based on A Combination of GLCM and LBP Features with PCA and IG

Application research of image recognition technology based on improved SVM in abnormal monitoring of rail fasteners

Railway transportation is the main means of transportation for people and the main way of logistics transportation, playing an important role in daily life. Therefore, the safety inspection of railway track has been widely valued. The abnormal intelligent detection of rail fasteners is the key content of rail safety detection. The traditional rail fastener detection method is based on machine learning for image recognition, such as SVM, to detect abnormal rail fasteners. But the traditional method has two defects. The first point is that the detection time is long, and the second point is that the detection accuracy is low. To solve this problem, a rail fastener anomaly detection model based on SVM optimized by IFOA algorithm is proposed. Firstly, the image of rail fastener is collected and filtered; Then, edge detection and image segmentation are performed to obtain the image of the target area; Finally, the HOG feature and LBP feature of the image are extracted, and the improved IFOA-SVM is used to recognize and classify the features, so as to achieve intelligent rail fastener anomaly detection. The experimental results show that when the IACO-SVM model is iterated to 254 times, the fitness value tends to be stable, which is 0.24. The detection accuracy of the model reaches 99.82%, which is higher than the traditional models, and can meet the work requirements of rail fastener anomaly detection. The rail fastener anomaly detection model based on SVM can improve the efficiency of rail fastener anomaly detection, and has a positive effect on the normal operation of railway transportation. However, the number of experimental samples used in the study is limited, which may lead to some errors in the experimental results. Therefore, it is necessary to increase the number of samples in subsequent studies.

Facial Expression Recognition Methods in the Wild Based on Fusion Feature of Attention Mechanism and LBP.

Facial expression methods play a vital role in human-computer interaction and other fields, but there are factors such as occlusion, illumination, and pose changes in wild facial recognition, as well as category imbalances between different datasets, that result in large variations in recognition rates and low accuracy rates for different categories of facial expression datasets. This study introduces RCL-Net, a method of recognizing wild facial expressions that is based on an attention mechanism and LBP feature fusion. The structure consists of two main branches, namely the ResNet-CBAM residual attention branch and the local binary feature (LBP) extraction branch (RCL-Net). First, by merging the residual network and hybrid attention mechanism, the residual attention network is presented to emphasize the local detail feature information of facial expressions; the significant characteristics of facial expressions are retrieved from both channel and spatial dimensions to build the residual attention classification model. Second, we present a locally improved residual network attention model. LBP features are introduced into the facial expression feature extraction stage in order to extract texture information on expression photographs in order to emphasize facial feature information and enhance the recognition accuracy of the model. Lastly, experimental validation is performed using the FER2013, FERPLUS, CK+, and RAF-DB datasets, and the experimental results demonstrate that the proposed method has superior generalization capability and robustness in the laboratory-controlled environment and field environment compared to the most recent experimental methods.

Lung Nodule Detection via Optimized Convolutional Neural Network: Impact of Improved Moth Flame Algorithm.

Lung cancer is a high-risk disease that affects people all over the world, and lung nodules are the most common signof early lung cancer. Since early identification of lung cancer can considerably improve a lung scanner patient's chances of survival, an accurate and efficient nodule detection system can be essential. Automatic lung nodule recognition decreases radiologists' effort, as well as the risk of misdiagnosis and missed diagnoses. Hence, this article developed a new lung nodule detection model with four stages like "Image pre-processing, segmentation, feature extraction and classification". In this processes, pre-processing is the first step, in which the input image is subjected to a series of operations. Then, the "Otsu Thresholding model" is used to segment the pre-processed pictures. Then in the third stage, the LBP features are retrieved that is then classified via optimized Convolutional Neural Network (CNN). In this, the activation function and convolutional layer count of CNN is optimally tuned via a proposed algorithm known as Improved Moth Flame Optimization (IMFO). At the end, the betterment of the scheme is validated by carrying out analysis in terms of certain measures. Especially, the accuracy of the proposed work is 6.85%, 2.91%, 1.75%, 0.73%, 1.83%, as well as 4.05% superior to the extant SVM, KNN, CNN, MFO, WTEEB as well as GWO + FRVM methods respectively.

LBP and CNN feature fusion for face anti-spoofing

LBP and CNN feature fusion for face anti-spoofing

Viscosity prediction using image processing and supervised learning

The paper implements image processing followed by supervised learning to predict the viscosity of hydrocarbon present at a specific reservoir depth in a heavy oil reservoir located in California, USA. Viscosity is predicted from white-light and UV-light images of side-wall core samples extracted from specific depths along the wellbore. The fluid viscosity primarily ranges from 5 to 600 cP at a temperature of 180 degF. The white-light and UV-light images are available for 680 side-wall core samples. Sobel, Sato, Hessian, LBP, and Multi-Otsu filters extract meaningful features from red, green, and blue pixel intensities of the white-light and UV-light images. The distribution of pixel-wise values for each filtered/raw image is further processed to derive histogram-based features. As a result, 600 features are extracted for each side-wall core sample. Thresholds based on the variance, MI score, F score, and Person’s r are imposed to select the most informative features for the viscosity prediction. Compared to filtered white-light images, raw RGB white-light images have stronger viscosity association. On the contrary, filtered UV-light images have stronger viscosity association as compared to raw RGB UV-light images. Supervised learning is deployed using the resulting features in both regression and classification of target viscosity. For the prediction of continuous-valued viscosity (regression), random forest is the best performer with a mean absolute error of 27 ∓ 3 cP. For the viscosity classification, high-viscosity and low-viscosity samples can be very well detected at a F1 Weighted Score higher than 0.95 and Matthew’s Correlation Score higher than 0.9. Overall, for viscosity prediction, classification is much superior to the regression. The novel image-based viscosity prediction workflow will help lower the cost and improve the precision of laboratory-based viscosity measurements and simultaneously contribute to a high-resolution reservoir model development.

ResNet18 Supported Inspection of Tuberculosis in Chest Radiographs With Integrated Deep, LBP, and DWT Features

ResNet18 Supported Inspection of Tuberculosis in Chest Radiographs With Integrated Deep, LBP, and DWT Features

Identification of cashmere and wool based on LBP and GLCM texture feature selection

There are invalid and redundant features in the texture feature extraction method of cashmere and wool fibers, which leads to the low recognition accuracy. In this paper, a novel texture feature selection method based on local binary pattern, the gray level co-occurrence matrix algorithm and chi-square test was proposed to sufficiently extract the effective features of these two fibers. Firstly, the collected images of cashmere and wool fibers are processed to obtain the clear texture images with background removed by pre-processing algorithm and local binary pattern. Then, the texture features are calculated by the gray level co-occurrence matrix, and the optimal 5-dimensional features are extracted by chi-square test to represent the texture information of cashmere and wool. Finally, the two fibers are automatically classified and recognized based on the support vector machine. The experimental results show that the proposed method obtained a high recognition accuracy with the percent of 94.39. It verifies that the method based on texture feature selection is effective to identify cashmere and wool fibers.

Performance comparison of various machine learning classifiers using a fusion of LBP, intensity and GLCM feature extraction techniques for thyroid nodules classification

Performance comparison of various machine learning classifiers using a fusion of LBP, intensity and GLCM feature extraction techniques for thyroid nodules classification

Content-Based Image Retrieval using Encoder based RGB and Texture Feature Fusion

Recent development of digital photography and the use of social media using smartphones has boosted the demand for image query by its visual semantics. Content-Based Image Retrieval (CBIR) is a well-identified research area in the domain of image and video data analysis. The major challenges of a CBIR system are (a) to derive the visual semantics of the query image and (b) to find all the similar images from the repository. The objective of this paper is to precisely define the visual semantics using hybrid feature vectors. In this paper, a CBIR system using encoded-based feature fusion is proposed. The CNN encoding features of the RGB channel are fused with the encoded texture features of LBP, CSLBP, and LDP separately. The retrieval performance of the different fused features is tested using three public datasets i.e. Corel-lK, Caltech, and 102flower. The result shows the class properties are better retained using the LDP with RGB encoded features, this helps to enhance the classification and retrieval performance for all three datasets. The average precision of Corel-lK is 94.5% and it is 89.7% for Caltech, and 88.7% for the 102flower. The average f1-score is 89.5% for Caltech, and 88.5% for the 102flower. The improvement in the f1-score value implies the proposed fused feature is more stable to deal the class imbalance problem.

Multiattention Joint Convolution Feature Representation With Lightweight Transformer for Hyperspectral Image Classification

Hyperspectral image (HSI) classification is currently a hot topic in the field of remote sensing. The goal is to utilize the spectral and spatial information from HSI to accurately identify land covers. Convolution neural network (CNN) is a powerful approach for HSI classification. However, CNN has limited ability to capture non-local information to represent complex features. Recently, vision transformers (ViTs) have gained attention due to their ability to process non-local information. Yet, under the HSI classification scenario with ultra-small sample rates, the spectral-spatial information given to ViTs for global modeling is insufficient, resulting in limited classification capability. Therefore, in this article, Multi-Attention Joint Convolution Feature Representation with Lightweight Transformer (MAR-LWFormer) is proposed, which effectively combines the spectral and spatial features of HSI to achieve efficient classification performance at ultra-small sample rates. Specifically, we use a three-branch network architecture to extract multi-scale convolved 3D-CNN, EMAP, and LBP features of HSI, respectively, by taking full exploitation of ultra-small training samples. Second, we design a series of multi-attention modules to enhance spectral-spatial representation for the three types of features and to improve the coupling and fusion of multiple features. Third, we propose an explicit feature attention tokenizer to transform the feature information, which maximizes the effective spectral-spatial information retained in the flat tokens. Finally, the generated tokens are input to the designed lightweight transformer for encoding and classification. Experimental results on three datasets validate that MAR-LWFormer has an excellent performance in HSI classification at ultra-small sample rates when compared to several state-of-the-art classifiers.

Joint transformer progressive self‐calibration network for low light enhancement

AbstractWhen the lighting conditions are poor and the environmental light is weak, the image captured by the imaging device often has lower brightness and is accompanied by a lot of noise. The paper designs a progressive self‐calibration network model (PSCNet) for recovering high‐quality low‐light‐enhanced images. First, shallow features in low‐light images can be better focused and extracted with the help of attention mechanism. Next, the feature mapping is passed to the encoder and decoder modules, where the transformer and encoder‐decoder jump connection structures can be better combined with the semantic information of the context to learn rich deep feature information. Finally, the self‐calibration module can adaptively cascade the features decoded by the decoder and input them into the residual attention module quickly and accurately. Meanwhile, the LBP features of the image are also fused into the feature information of the residual attention module to enhance the detailed texture information of the image. Qualitative analysis and quantitative comparison of a large number of experimental results show that this method outperforms existing methods.

Automated Detection of Multitype Landforms on Mars Using a Light-Weight Deep Learning-Based Detector

Intification of geological salient landforms is a primary requirement for spacecraft motion estimation and obstacle avoidance. As a large volume of high-resolution images are acquired by the Mars reconnaissance orbiter, growing number of approaches are proposed to develop automated approaches to detect a particular landform on Mars. However, most existing objective detection models are limited to sliding window-based and morphology-based algorithms, which require complicated preprocessing operations and can hardly be generalized to detecting different types of landforms. In this article, we aimed at developing a multitype landform detection system based on a light-weight deep learning framework, which has a quite small model size but presents excellent performance. This specific deep learning-based framework is named as mini shot multibox detector (SSD), by downsizing and modifying the existing single SSD. In the mini-SSD, some components are further optimized to adapt to this domain specific problem. A pretraining strategy is well-designed and merged into the entire model training process. In the performance evaluation tests, the proposed framework was trained and tested on images with different scales collected from different locations in high resolution imaging science experiment database. Results demonstrate that the introduced Adam optimizer and pretraining strategy can form positive effective to both model training and inference performance. The proposed framework and strategy combination outperforms the original SSD300, faster R-CNN, and YOLO series models as well as all state-of-the-art sliding window-based detectors in the field, namely AdaBoost with LBP features, AdaBoost with Haar features, and support vector machines with histogram of oriented gradient (HOG) features, in two different testing sets. Additionally, it shows high resilience on detecting target landforms in different environments with various sizes and shapes from the qualitative analysis, and can be generalized as a tool for relevant applications.

Segmentation of Retinal Blood Vessels Using U-Net++ Architecture and Disease Prediction

This study presents a segmentation method for the blood vessels and provides a method for disease diagnosis in individuals based on retinal images. Blood vessel segmentation in images of the retina is very challenging in medical analysis and diagnosis. It is an essential tool for a wide range of medical diagnoses. After segmentation and binary image improvement operations, the resulting binary images are processed and the features in the blood vessels are used as feature vectors to categorize retinal images and diagnose the type of disease available. To carry out the segmentation task and disease diagnosis, we used a deep learning approach involving a convolutional neural network (CNN) and U-Net++ architecture. A multi-stage method is used in this study to better diagnose the disease using retinal images. Our proposed method includes improving the color image of the retina, applying the Gabor filter to produce images derived from the green channel, segmenting the green channel by receiving images produced from the Gabor filter using U-Net++, extracting HOG and LBP features from binary images, and finally disease diagnosis using a one-dimensional convolutional neural network. The DRIVE and MESSIDOR image banks have been used to segment the image, determine the areas related to blood vessels in the retinal image, and evaluate the proposed method for retinal disease diagnosis. The achieved results for accuracy, sensitivity, specificity, and F1-score are 98.9, 94.1, 98.8, 85.26, and, 98.14, respectively, in the DRIVE dataset and the obtained results for accuracy, sensitivity, and specificity are 98.6, 99, 98, respectively, in MESSIDOR dataset. Hence, the presented system outperforms the manual approach applied by skilled ophthalmologists.

Ripeness Identification of Chayote Fruits using HSI and LBP Feature Extraction with KNN Classification

This study aims to build a system to identify the ripeness level of chayote that can be done easily and without damaging the quality of the chayote. This study employs digital image processing technology using Hue Saturation Intensity color feature extraction and texture feature extraction of Local Binary Pattern with K-Nearest Neighbor classification so that the process of identifying the ripeness level of chayote will be easier and more effective. This study uses 100 image datasets and is carried out by taking photos of chayote. The stages in this study include the input of chayote images followed by the image pre-processing stage. Next is feature extraction which is divided into three scenarios, namely HSI feature extraction, LBP feature extraction and a combination of the two feature extractions. The final stage is to classify objects that are closest to the object being tested using the KNN method. By determining the value of K in the KNN classification method, the results show that the use of the Chebyshev distance calculation model in LBP feature extraction with K = 5 is a test that has the best accuracy of 90%.

Multi-Method Diagnosis of CT Images for Rapid Detection of Intracranial Hemorrhages Based on Deep and Hybrid Learning

Intracranial hemorrhaging is considered a type of disease that affects the brain and is very dangerous, with high-mortality cases if there is no rapid diagnosis and prompt treatment. CT images are one of the most important methods of diagnosing intracranial hemorrhages. CT images contain huge amounts of information, requiring a lot of experience and taking a long time for proper analysis and diagnosis. Thus, artificial intelligence techniques provide an automatic mechanism for evaluating CT images to make a diagnosis with high accuracy and help radiologists make their diagnostic decisions. In this study, CT images for rapid detection of intracranial hemorrhages are diagnosed by three proposed systems with various methodologies and materials, where each system contains more than one network. The first system is proposed by three pretrained deep learning models, which are GoogLeNet, ResNet-50 and AlexNet. The second proposed system using a hybrid technology consists of two parts: the first part is the GoogLeNet, ResNet-50 and AlexNet models for extracting feature maps, while the second part is the SVM algorithm for classifying feature maps. The third proposed system uses artificial neural networks (ANNs) based on the features of the GoogLeNet, ResNet-50 and AlexNet models, whose dimensions are reduced by a principal component analysis (PCA) algorithm, and then the low-dimensional features are combined with the features of the GLCM and LBP algorithms. All the proposed systems achieved promising results in the diagnosis of CT images for the rapid detection of intracranial hemorrhages. The ANN network based on fusion of the deep feature of AlexNet with the features of GLCM and LBP reached an accuracy of 99.3%, precision of 99.36%, sensitivity of 99.5%, specificity of 99.57% and AUC of 99.84%.

Application of Mobile Virtual Reality Technology Combined with Neural Network in Facial Expression Recognition.

In recent years, in the field of virtual reality, in more and more scenes, users interact with hardware or programs through facial expressions. In order to give full play to the advantages of program interaction between virtual reality devices and users, this paper proposes a mobile virtual reality expression recognition system combined with convolution neural network. Based on the optimized AlexNet network, an expression recognition algorithm is constructed and combined with LBP feature mapping technology to improve the performance of the algorithm. At the same time, according to the nature and characteristics of mobile virtual reality devices, the user face information acquisition algorithm is optimized. The performance test results of the expression recognition system show that the recognition accuracy of the system is higher than that of the traditional convolution neural network expression recognition algorithm, and the maximum difference is greater than 10%. At the same time, the average running speed of the whole system is about 37 ms, which can meet the accuracy and real-time requirements of expression recognition in virtual reality interaction. The experimental results show that the expression recognition system proposed in this paper can be applied to mobile virtual reality devices. At the same time, it also provides new ideas for industry researchers to optimize the identification function.

Neutrosophic set based local binary pattern for texture classification

This paper presents novel neutrosophic set-based Completed Local Binary Pattern (CLBP) hybrid methods. These methods first transform the input image into a neutrosophic domain, and the image's texture is characterized by truth and false membership components. Grayscale images are more robust to noise in the neutrosophic domain. The neutrosophic components suppress noise, by this way edges in the images may be detected more precisely. As a result, using neutrosophic truth and false sets with a grayscale input image has resulted in more robust features. Besides, these discriminative features are combined with rotation invariant LBP features with a cross-scale joint coding strategy. The proposed methods can contribute to the classification performance with a reasonable computational cost. When compared to state-of-the-art hand-crafted approaches, experimental findings on state-of-the-art texture databases show that the proposed methods can enhance classification accuracy by roughly 24%. Moreover, the proposed method achieves up to 34% better results than the state-of-the-art deep learning-based methods.

Window averaging to create LBP features for a color image

Digital color images are widely used in various computer applications. Some of these applications such as digital image recognition systems require dealing with image classifiers which can be formed as a set of values representing the image. In this paper a simple method of image features victor creation will be proposed, the method will be tested and implemented, the obtained experimental results will be analyzed to ensure that each generated features victor is unique. The proposed method will be based on a window averaging, the window size can be easily changed without affecting the features victor size.