Histogram Of Oriented Gradients Research Articles

Rapid and Precise Identification of COVID-19 through Segmentation and Classification of CT and X-ray Images

The COVID-19 pandemic, caused by a novel coronavirus, has become a global epidemic. Although the reverse transcription-polymerase chain reaction (RT-PCR) test is the current gold standard for detecting the virus, its low reliability has led to the use of CT and X-ray imaging in diagnostics. As limited vaccine availability necessitates rapid and accurate detection, this study applies k-means and fuzzy c-means segmentation to CT and X-ray images to classify COVID-19 cases as either diseased or healthy for CT scans and diseased, healthy, or non-COVID pneumonia for X-rays. Our research employs four open-access, widely-used datasets and is conducted in four stages: preprocessing, segmentation, feature extraction, and classification. During feature extraction, we employ the Gray-Level Co-Occurrence Matrix (GLCM), Local Binary Pattern (LBP), and Histogram of Oriented Gradients (HOG). In the classification process, our approach involves utilizing k-Nearest Neighbor (kNN), Support Vector Machines (SVM), and Extreme Learning Machines (ELM) techniques. Our research achieved a sensitivity rate exceeding 99%, which is higher than the 60-70% sensitivity rate of PCR tests. As a result, our study can serve as a decision support system that can help medical professionals make rapid and precise diagnoses with a high level of sensitivity.

Automated Grading System for Breast Cancer Histopathological Images Using Histogram of Oriented Gradients (HOG) Algorithm

Breast cancer is the most common type of cancer in the world, affecting both men and women. In 2023, the American Cancer Society's reported that there will be approximately 297,800 new cases of invasive breast cancer in women and 2,850 in men, along with 55,750 cases of ductal carcinoma in situ (DCIS) in women. Further, an estimated 43,750 deaths are expected from breast cancer, of which approximately 43,180 are among women and 570 are among men. In this paper, we propose an automated grading system for breast cancer based on tumor's histopathological images using a combination of the Histogram of Oriented Gradients (HOG) for feature extraction and machine learning algorithms. The proposed system has four main phases: image preprocessing and segmentation, feature extraction, classification, and integration with a website. Grayscale conversion, enhancement, noise and artifact removal methods are used during the image preprocessing stage. Then the image is segment during the segmentation phase to extract regions of interest. And then, features are extracted from the obtained region of interest using the Histogram of Oriented Gradients (HOG) algorithm. The next, the images are classified into three distinct breast cancer grades based on the extracted features using machine learning algorithms. Moreover, the effectiveness of the proposed system was evaluated and reported using vary evaluation methods and the results showed a remarkable accuracy of up to 97% by the SVM classifier. Finally, the machine learning model is integrated into a website to improve the detection and diagnosis of breast cancer disease and facilitate the access and use of patient data. This will make the work easier for physicians to enhance breast cancer detection and treatment

Next-generation attendance tracking: Automated face detection for lecture attendance

Face detection technology has garnered much interest due to its potential uses in industries like face identification and video surveillance systems. In addition to being a component of the automatic face recognition framework, real-time face detection is now becoming its study area. There are numerous methods to handle the problem of face detection. Traditional methods of manual attendance tracking in lecture halls are time-consuming, prone to errors, and lack efficiency. Next-generation attendance tracking systems are incorporating automated face detection technology to address these challenges. This paper explores the implementation of automated face detection for lecture attendance, leveraging computer vision and facial recognition algorithms. Initially, students’ data were gathered and preprocessed using a Median filter (MF). Then the preprocessed data features were extracted using a histogram of gradients (HOG). And finally, extracted features were classified using a Backpropagation Bayesian neural network (BBNN). To prove the efficiency of the suggested method, that is compared and contrasted with specific conventional methods. Experimental outcomes show that the recommended method performs superior when contrasted to traditional methods.

TwSense: Highly Robust Through-the-Wall Human Detection Method Based on COTS Wi-Fi Device

With the popularization of Wi-Fi router devices, the application of device-free sensing has garnered significant attention due to its potential to make our lives more convenient. Wi-Fi signal-based through-the-wall human detection offers practical applications, such as emergency rescue and elderly monitoring. However, the accuracy of through-the-wall human detection is hindered by signal attenuation caused by wall materials and multiple propagation paths of interference. Therefore, through-the-wall human detection presents a substantial challenge. In this paper, we proposed a highly robust through-the-wall human detection method based on a commercial Wi-Fi device (TwSense). To mitigate interference from wall materials and other environmental factors, we employed the robust principal component analysis (OR-PCA) method to extract the target signal of Channel State Information (CSI). Subsequently, we segmented the action-induced Doppler shift feature image using the K-means clustering method. The features of the images were extracted using the Histogram of Oriented Gradients (HOG) algorithm. Finally, these features were fed into an SVM classifier (G-SVM) optimized by a grid search algorithm for action classification and recognition, thereby enhancing human detection accuracy. We evaluated the robustness of the entire system. The experimental results demonstrated that TwSense achieved the highest accuracy of 96%.

Hybrid feature extraction technique for automatic classification of COVID-19 chest CT images

ABSTRACT COVID-19 has seriously affected normal life as well as public safety. It is extremely transmissible and has now infected millions of people worldwide. To obtain more image features of the lungs, Computed Tomography (CT) scans are widely used. However, manual examination of CT images for abnormal areas of COVID-19 disease can be time-consuming, and it is highly subjective to determine whether they are infected. To rapidly screen patients, Machine Learning (ML) can be used to determine whether patients have the disease. In this paper, a hybrid extraction technique is used to extract feature vectors from CT images, which is a mixture of a histogram of orientation gradients (HOG) extraction technique and a local binary pattern (LBP) extraction technique. In this experiment, 960 NON-COVID-19 and 960 COVID-19 were adopted to train the model, and 240 NON-COVID-19 and 240 COVID-19 were used to test the model. And the CT images were scaled to a uniform size. After obtaining the feature vectors using HOG and LBP feature extraction methods, The CT images were classified using a Support Vector Machine (SVM) classifier optimised by Particle Swarm Optimisation (PSO). In the performance evaluation of the presented classification model, the combination of the HOG feature extraction technique and the LBP feature extraction technique resulted in a substantial improvement in the classification effectiveness of the SVM. HOG_LBP PSO SVM improved Accuracy to 97.5%, Precision to 97.75%, Recall to 97.27%, Specificity to 97.25%, F1_score to 97.50%, and Mcc to 95.01%.

Design of supervision-scalable learning systems: Methodology and performance benchmarking

The design of robust learning systems that offer stable performance under a wide range of supervision degrees is investigated in this work. We choose the image classification problem as an illustrative example and focus on the design of modularized systems that consist of three learning modules: representation learning, feature learning, and decision learning. We discuss ways to adjust each module so that the design is robust with respect to different training sample numbers. Based on these ideas, we propose two families of learning systems. One adopts the classical histogram of oriented gradients (HOG) features, while the other uses successive-subspace-learning (SSL) features. We test their performance against LeNet-5 and ResNet-18, two end-to-end optimized neural networks, for MNIST, Fashion-MNIST and CIFAR-10 datasets. The number of training samples per image class goes from the extremely weak supervision condition (i.e., one labeled sample per class) to the strong supervision condition (i.e., 4096 labeled samples per class) with a gradual transition in between (i.e., 2n, n=0,1,…,12). Experimental results show that the two families of modularized learning systems have more robust performance than LeNet-5 and ResNet-18. They both outperform the two deep learning networks by a large margin for small n and have performance comparable for large n.

A novel framework using binary attention mechanism based deep convolution neural network for face emotion recognition

The use of facial emotion recognition (FER) technologies will become more pervasive in our everyday lives. Emotional awareness is advantageous for many types of businesses and areas of life. It is advantageous and important for reasons of security and heath. Deep hierarchical FER systems often focus on the following two main problems; going out of control due to identification factors including lighting, face location, and recognition bias, as well as a lack of training data. We developed each Deep Convolutional Neural Network (DCNN) based on a Binary Attention Mechanism (BAM) for the facial emotion recognition issue in our proposed system. Each image of a face has to be assigned to one of the seven facial emotions. An updated BAM-DCNN model was trained using the original pixel data characteristics. The Histogram of Oriented Gradients (HOG) is used for data preparation. To lessen the overfitting of the models, we used dropout and batch normalization in addition to L2 regularization. The recommended technique enables the detection of human emotion in images by automatically recognizing, extracting, and evaluating diverse face expressions. We extract and examine performance assessment measures from FER datasets, including recognition accuracy, precision, sensitivity, specificity, recall, and F1 score. To demonstrate the effectiveness of our system, we also contrast the recommended technique with the practices now in use.

Breast cancer classification using deep learned features boosted with handcrafted features

Breast cancer is one of the leading causes of death among women across the globe. It is difficult to treat if detected at advanced stages. However, early detection can significantly increase chances of survival and improves the lives of millions of women. Given the widespread prevalence of breast cancer, it is of utmost importance for the research community to provide comprehensive framework encompassing early detection, classification, and diagnosis. The artificial intelligence research community, in coordination with medical practitioners, is developing such frameworks to automate the task of detection. With the surge in research activities coupled with the availability of large datasets and enhanced computational powers, it is expected that AI framework results will help even more clinicians in making correct predictions. In this article, a novel framework for the classification of breast cancer using mammograms is proposed. The proposed framework not only uses features extracted from Convolutional Neural Network (CNN) but also combines these features with handcrafted features (Histogram of Oriented Gradients (HOG) and Local Binary Pattern (LBP)) which helps in embedding domain expert knowledge, a step towards making proposed framework transparent. Experimental results conducted on the CBIS-DDSM dataset demonstrate that the proposed framework outperforms the current state-of-the-art methods in breast cancer classification.

AGE and GENDER PREDICTON from HAND WRITTING TEXT using ATTENTION-BASED CNN

The goal is to predict the age range and gender of a person based on their handwriting. Two feature extraction methods, Histogram of Oriented Gradients and Linear Discriminant Analysis, are employed, along with two machine learning techniques, Support Vector Machine and Attention Based Convolutional Neural Network. For each feature extraction method, the system predicts the age range and gender using SVM and ABCNN. The system provides accuracy metrics for each prediction to decide which technique is best, indicating the reliability of the age and gender estimations. This is designed to work in real-time, enabling quick predictions for incoming handwriting samples. Key Words : Histogram of Oriented Gradients (HOG), Linear Discriminant Analysis (LDA), Support Vector Machine (SVM), Attention Based Convolutional Neural Network (ABCNN) , Convolutional Neural Network (CNN)

Multi-scale feature fusion-based lightweight dual stream transformer for detection of paddy leaf disease.

Traditionally, rice leaf disease identification relies on a visual examination of abnormalities or an analytical result obtained by growing bacteria in the research lab. This method of visual evaluation is qualitative and error-prone. On the other hand, an artificial neural network system is fast and more accurate. Several pieces of research using traditional machine learning and deep convolution neural networks (CNN) have been utilized to overcome the issues. Still, these methods need more semantic contextual global and local feature extraction. Due to this, efficiency is less. Hence, in the present study, a multi-scale feature fusion-based RDTNet has been designed. The RDTNet contains two modules, and the first module extracts feature via three scales from the local binary pattern (LBP), gray, and a histogram of orient gradient (HOG) image. The second module extracts semantic global and local features through the transformer and convolution block. Furthermore, the computing cost is reduced by dividing the query into two parts and feeding them to convolution and the transformer block. The results indicate that the proposed method has a very high average precision, f1-score, and accuracy of 99.55%, 99.54%, and 99.53%, respectively. It is suggestive of improved classification accuracy using multi-scale features and the transformer. The model has also been validated on other datasets confirming that the present model can be used for real-time rice disease diagnosis. In the future, such models can be used for monitoring other crops, including wheat, tomato, and potato.

Identification of Solid and Liquid Materials Using Acoustic Signals and Frequency-Graph Features.

Material identification is playing an increasingly important role in various sectors such as industry, petrochemical, mining, and in our daily lives. In recent years, material identification has been utilized for security checks, waste sorting, etc. However, current methods for identifying materials require direct contact with the target and specialized equipment that can be costly, bulky, and not easily portable. Past proposals for addressing this limitation relied on non-contact material identification methods, such as Wi-Fi-based and radar-based material identification methods, which can identify materials with high accuracy without physical contact; however, they are not easily integrated into portable devices. This paper introduces a novel non-contact material identification based on acoustic signals. Different from previous work, our design leverages the built-in microphone and speaker of smartphones as the transceiver to identify target materials. The fundamental idea of our design is that acoustic signals, when propagated through different materials, reach the receiver via multiple paths, producing distinct multipath profiles. These profiles can serve as fingerprints for material identification. We captured and extracted them using acoustic signals, calculated channel impulse response (CIR) measurements, and then extracted image features from the time-frequency domain feature graphs, including histogram of oriented gradient (HOG) and gray-level co-occurrence matrix (GLCM) image features. Furthermore, we adopted the error-correcting output code (ECOC) learning method combined with the majority voting method to identify target materials. We built a prototype for this paper using three mobile phones based on the Android platform. The results from three different solid and liquid materials in varied multipath environments reveal that our design can achieve average identification accuracies of 90% and 97%.

A Fine-Tuned Hybrid Stacked CNN to Improve Bengali Handwritten Digit Recognition

Recognition of Bengali handwritten digits has several unique challenges, including the variation in writing styles, the different shapes and sizes of digits, the varying levels of noise, and the distortion in the images. Despite significant improvements, there is still room for further improvement in the recognition rate. By building datasets and developing models, researchers can advance state-of-the-art support, which can have important implications for various domains. In this paper, we introduce a new dataset of 5440 handwritten Bengali digit images acquired from a Bangladeshi University that is now publicly available. Both conventional machine learning and CNN models were used to evaluate the task. To begin, we scrutinized the results of the ML model used after integrating three image feature descriptors, namely Binary Pattern (LBP), Complete Local Binary Pattern (CLBP), and Histogram of Oriented Gradients (HOG), using principal component analysis (PCA), which explained 95% of the variation in these descriptors. Then, via a fine-tuning approach, we designed three customized CNN models and their stack to recognize Bengali handwritten digits. On handcrafted image features, the XGBoost classifier achieved the best accuracy at 85.29%, an ROC AUC score of 98.67%, and precision, recall, and F1 scores ranging from 85.08% to 85.18%, indicating that there was still room for improvement. On our own data, the proposed customized CNN models and their stack model surpassed all other models, reaching a 99.66% training accuracy and a 97.57% testing accuracy. In addition, to robustify our proposed CNN model, we used another dataset of Bengali handwritten digits obtained from the Kaggle repository. Our stack CNN model provided remarkable performance. It obtained a training accuracy of 99.26% and an almost equally remarkable testing accuracy of 96.14%. Without any rigorous image preprocessing, fewer epochs, and less computation time, our proposed CNN model performed the best and proved the most resilient throughout all of the datasets, which solidified its position at the forefront of the field.

SDPS-40 MEDULLOBLASTOMA MOLECULAR SUBTYPE PREDICTION USING PATHOMIC FEATURE ANALYSIS

Abstract The clinical course of pediatric medulloblastoma patients exhibits wide variability based on molecular subtype. The ability to predict medulloblastoma molecular subtypes (Group 3/4, WNT, SHH) based on histopathologic features alone has immense utility by providing clinicians insight into the molecular and biological underpinnings of the tumor. Although the use of computational pathology (“pathomics”) is expanding in multiple disciplines, its application in pediatric neuro-oncology has been limited. Our preliminary analysis highlights the potential for application in medulloblastoma and pediatric brain tumor research. Using the Children’s Brain Tumor Network database, pediatric patients with available clinical data, whole-slide images (WSI) and molecular subtyping were evaluated for inclusion. Regions-of-interest (ROI) on each H&E-stained WSI were selected for pathomic analysis, and features were extracted using a digital pathology software package (Qupath). 25 pathomic features capturing cellular and nuclear-level intensity and morphology properties were included. Subsequently, a histogram gradient boosting classifier was used to perform binary classification of Group 3/4 vs. SHH/WNT with leave-one-subject-out cross validation. A total of 68 subjects, with median age at diagnosis of 8.19 years (range 0.29-18.69), and median overall survival (OS) of 56 months (range 0.76-258.37) were identified. Our cohort included 44 and 24 subjects in Group 3/4 and SHH/WNT, respectively. Pathomic features classified subjects into Group 3/4 or SHH/WNT with a classification accuracy of 71% (SEM=0.06). A t-test against stratified chance accuracy (57%) showed significantly above-chance testing performance (p=.019). Inspection of the confusion matrix indicated better discrimination of Group 3/4 compared to SHH/WNT subjects, likely due to the larger number of these subjects in the dataset. Our preliminary analysis shows acceptable testing accuracy in distinguishing Group 3/4 from SHH/WNT medulloblastoma, and highlights the possibility of gaining molecular insight purely from histopathologic images. The potential of pathomic features should be considered an untapped resource.

PFP-HOG: Pyramid and Fixed-Size Patch-Based HOG Technique for Automated Brain Abnormality Classification with MRI.

Detecting neurological abnormalities such as brain tumors and Alzheimer's disease (AD) using magnetic resonance imaging (MRI) images is an important research topic in the literature. Numerous machine learning models have been used to detect brain abnormalities accurately. This study addresses the problem of detecting neurological abnormalities in MRI. The motivation behind this problem lies in the need for accurate and efficient methods to assist neurologists in the diagnosis of these disorders. In addition, many deep learning techniques have been applied to MRI to develop accurate brain abnormality detection models, but these networks have high time complexity. Hence, a novel hand-modeled feature-based learning network is presented to reduce the time complexity and obtain high classification performance. The model proposed in this work uses a new feature generation architecture named pyramid and fixed-size patch (PFP). The main aim of the proposed PFP structure is to attain high classification performance using essential feature extractors with both multilevel and local features. Furthermore, the PFP feature extractor generates low- and high-level features using a handcrafted extractor. To obtain the high discriminative feature extraction ability of the PFP, we have used histogram-oriented gradients (HOG); hence, it is named PFP-HOG. Furthermore, the iterative Chi2 (IChi2) is utilized to choose the clinically significant features. Finally, the k-nearest neighbors (kNN) with tenfold cross-validation is used for automated classification. Four MRI neurological databases (AD dataset, brain tumor dataset 1, brain tumor dataset 2, and merged dataset) have been utilized to develop our model. PFP-HOG and IChi2-based models attained 100%, 94.98%, 98.19%, and 97.80% using the AD dataset, brain tumor dataset1, brain tumor dataset 2, and merged brain MRI dataset, respectively. These findings not only provide an accurate and robust classification of various neurological disorders using MRI but also hold the potential to assist neurologists in validating manual MRI brain abnormality screening.

Sea Cucumber Identification Fishing System Based on Hog Super Pixel Feature Extraction

Due to the refraction, reflection and absorption of light induced by seawater and suspended particles, the image recognition of underwater vehicle has the problem of low accuracy and difficult localization, and to solve this one problem, an improved LSTM classification algorithm (CH-LSTM) based on CLAHE and HOG is given. First, the improved CLAHE algorithm is used to enhance image by image noise removal and de-fogging, the problems of underwater image distortion, fine lines, and abrupt changes are solved in the aquaculture waters image. Then, the Histogram of Orientation Gradient (HOG) operator is used for feature extraction to describe the shape of sea cucumbers and generate feature vectors; finally, the LSTM algorithm is used to classify the feature vectors and avoid overfitting through memory gates, so the generalization ability of the model was enhanced. The identification and sorting experiments of underwater fishing sea cucumbers show that the proposed algorithm is superior, especially in the environment of poor water quality, and the identification and localization accuracy of sea cucumbers is improved compared with the traditional ground SVM and BP neural network algorithms, and the identification accuracy is above 95.28%.

Classification Size of Underground Object from Ground Penetrating Radar Image using Machine Learning Technique

Ground Penetrating Radar (GPR) is a useful tool in detecting subsurface object or hidden structure defects However, the time-consuming problems and high requirement of professional manpower is required to analyse the GPR data. Machine learning is a tool that endowed with the ability to learn, and it can reduce time taken for the GPR data analysing. To simplify the identification process, a framework is proposed to classify the size of underground metallic pipe by using Histogram of Oriented Gradient (HOG) as a feature extraction algorithm. Two machine learning algorithms namely Support Vector Machines (SVM) and Backpropagation Neural Network were proposed to classify the size of the underground metallic pipe. As a result, the accuracy from the identification is more than 98% for both classifier algorithm.

An adaptive system for predicting student attentiveness in online classrooms

The effectiveness of teaching methods is determined by whether a student is attentive in a lecture or not. In face-to-face classroom teaching, a teacher is able to judge whether students are understanding the subject, based on their facial expressions. However, since the uprise of COVID-19 pandemic, virtual classrooms have found a holding in the field of education and detecting attentiveness of students is a challenge in the same. This paper proposes a student attentiveness model that would detect and monitor a student’s eye state to determine their level of attentiveness and provide a real-time feedback mechanism to the teacher. The proposed model employs a histogram of oriented gradient (HOG) method in conjunction with support vector machine (SVM) algorithm for face recognition. It then computes an adaptive eye aspect ratio (AEAR) for each individual student to determine their level of attentiveness. The model is tested on a real-time dataset and validated using classifiers (SVM, decision tree, and random forest). The results of the classifiers verify that the model produces an accuracy of more than 92%.

Brain tumor Classification using Optimized and Relief-based Feature Reduction and Regression Neural Network

The brain is an essential part of the human body for carrying out normal functions. Nowadays, a person's fast-paced lifestyle combined with an unhealthy diet can lead to brain damage. The current state of medicine and imaging could prove to be a great boon in this case. The image's restricted borders and lack of textural variety make it difficult to detect the tumour even in its earliest stages. Before, a picture was categorised with the help of a hybrid feature extraction system and an Extreme Learning Machine. Science also makes use of projection-based categorization. In this study, we will use MRI image characteristics to categorize different types of brain cancer. This is how a neural network is constructed based on patterns for fast processing. An anisotropic diffusion filter was used in the preprocessing of these images. The cancerous tissue is then separated using thresholding and a morphological strategy. Based on the local binary patterns and histogram-oriented gradients of cancer areas, the pattern-oriented features were generated. What we care about are the global, first-order properties of the image. In order to better all learned characteristics, the error rate and the method of relief are both decreased. A regression neural network is then used to categories the feature with the highest ratings from both PSO and the relief technique. All of the work is done in MATLAB R2020b on Windows 10 with the fig-sharing tumour dataset. Finally, its accuracy, sensitivity, and specificity are compared to those of alternative methods.

Performance Analysis of Naive Bayes Variation Method in Spice Image Classification Using Histogram of Gradient Oriented (HOG) Feature Extraction

Indonesia has a lot of natural wealth of spices. The diversity of spices is an inseparable aspect of Indonesian history. Spices and seasonings are biological resources that have long played an important role in human life. Indonesian spices have almost the same color and shape. The purpose of this study was to analyze the performance of the Naïve Bayes variation method in classifying spices using a Histogram Of Oriented Gradient (HOG) feature extraction. Based on 3 tests, the performance of the four Naïve Bayes variation methods carried out in this study, it can be seen that testing 5 types of spices using the Gaussian Naïve Bayes method obtained the best performance with an accuracy of 0.946, a precision of 0.95, a recall of 0.945, f1 score of 0.947, f beta score of 0.946, and Jaccard score of 0.90. Where as using the Complement Naïve Bayes method gets the lowest performance. From the results of this study it can be concluded that by utilizing HOG feature extraction and the Naïve Bayes variation method, maximum classification results are obtained in classifying spices. To obtain more accurate classification results, consider using other methods and other feature extraction

Highly effective end-to-end single-to-multichannel feature fusion and ensemble classification to decode emotional secretes from small-scale spontaneous facial micro-expressions

Facial Micro-Expression (ME) is one of the pre-dominating non-verbal clues to demystify the true emotional states that people try to conceal cautiously. But emotion recognition from spontaneous ME images limits the high accuracy due to short duration and low intensity, lack of sufficient samples and consistencies among publicly available ME datasets. In this study, we have proposed two highly effective, lightweight, and generalized single-channel DLRRF-MER, and multi-channel DLH-3C-FUSION fusion models inspired by deep dense convolutional models and texture-based feature descriptors Local Binary Pattern (LBP) and Histogram of Oriented Gradients (HOG) to recognize ME from apex frame by constructing a composite dataset from five publicly available ME datasets CASME, CASMEII, CAS(ME)2, SAMM and MMEW. Pre-training has been done on a new composition of five facial macro expressions datasets CK+, MUGFE, OuluCasia, SFEW, and RAF-DB. The proposed models are fine-tuned on the target ME dataset rigorously with Stratified 5-Fold and 10-Fold, Leave-One-Subject-Out, and Leave-One-Dataset-Out(LODO) cross-validations(CV). In all evaluations, our proposed algorithms show remarkable improvement in effectiveness which surpasses the state-of-the-art accuracies and results in higher generalization capacity.