Basis Kernel Research Articles

Predicting Driver Errors during Automated Vehicle Takeovers

The transition of control between partially automated vehicles and drivers is an important part of the operational design domain and poses unique and important design issues. One approach for enhancing the design of the transition of control mechanisms is to predict driver behavior during a takeover by analyzing his/her state before a takeover. Although there is a wealth of existing literature on modeling the prediction of driver behavior, little is known about the prediction of takeover performance (e.g., driver error) and its underlying data structure (e.g., window sizes or the inclusion of certain features). Thus, the goal of this study is to predict driver error during a takeover event using supervised machine learning algorithms for various window sizes. Three machine learning algorithms (i.e., decision tree, random forest, and support vector machine with a radial basis kernel) were applied to granular driving performance, physiological, and glance data from a driving simulator experiment examining automated vehicle driving. The results showed that a random forest algorithm with an area under the receiver operating characteristic curve of 0.72, trained on a 3 s window before the takeover time, had the best performance with regard to classifying driver error accurately. In addition, we identified the 10 most important predictors that resulted in the best error prediction performance. The results of this study could be useful in developing algorithms for driver state that could be integrated into highly automated systems and, potentially, improve the takeover process.

Influence of elevation on plastic-covered greenhouse mapping accuracy using high-resolution satellite imagery and machine learning algorithms

ABSTRACT Proliferation of plastic-covered greenhouse (PCG) farming areas has been unparalleled in recent decades around the globe. Therefore, PCG’s proper mapping and continuous monitoring are required to maintain protected farming sustainability and to reduce plastic pollution. Topography is one of the key factors affecting PCG mapping accuracy from satellite imagery. In this research, in order to understand the influence of elevation on PCG mapping accuracy in a Moroccan (with altitude below 200 m) and Vietnamese (an average altitude of 1,500 m) area, five supervised algorithms were applied to high-resolution PlanetScope and RapidEye imagery: maximum likelihood classification (MLC), k-nearest neighbours (k-NN), random forest (RF), support vector machine with linear kernel (SVML) and radial basis kernel (SVMRB). The findings indicated that k-NN (SVMRB) produced the greatest overall accuracy, followed by SVMRB (RF) for Dalat City (Loukkos). The present study also found that RapidEye outperforms PlanetScope in both areas and independently of the machine learning algorithm. Elevation data inclusion significantly improves (1.6% to 6.8%) various satellites’ performance, particularly in Dalat City. Thus, this study demonstrated that topography has an impact on PCG mapping, especially in hilly areas.

SimpleMKKM: Simple Multiple Kernel K-Means.

We propose a simple yet effective multiple kernel clustering algorithm, termed simple multiple kernel k-means (SimpleMKKM). It extends the widely used supervised kernel alignment criterion to multi-kernel clustering. Our criterion is given by an intractable minimization-maximization problem in the kernel coefficient and clustering partition matrix. To optimize it, we equivalently rewrite the minimization-maximization formulation as a minimization of an optimal value function, prove its differenentiablity, and design a reduced gradient descent algorithm to decrease it. Furthermore, we prove that the resultant solution of SimpleMKKM is the global optimum. We theoretically analyze the performance of SimpleMKKM in terms of its clustering generalization error. After that, we develop extensive experiments to investigate the proposed SimpleMKKM from the perspectives of clustering accuracy, advantage on the formulation and optimization, variation of the learned consensus clustering matrix with iterations, clustering performance with varied number of samples and base kernels, analysis of the learned kernel weight, the running time and the global convergence. The experimental study demonstrates the effectiveness of the proposed SimpleMKKM by considerably and consistently outperforming state of the art multiple kernel clustering alternatives. In addition, the ablation study shows that the improved clustering performance is contributed by both the novel formulation and new optimization. Our work provides a more effective approach to integrate multi-view data for clustering, and this could trigger novel research on multiple kernel clustering. The source code and data for SimpleMKKM are available at https://github.com/xinwangliu/SimpleMKKMcodes/.

Validate Model Endorsed for Support Vector Machine Alignment with Kernel Function and Depth Concept to Get Superlative Accurateness

A support vector machine (SVM) is authoritative tool for statistical learning model which is well proved based on the literature reviews which is rooted in finding the operational risk. The Key factor is kernel function and its parameters selection. Once the debate of finalizing the influence factor (i.e) kernel parameters and error penalty factors, we can able to find the new kernel function as a proposed model by bring together the kernel with robust depth procedures. Here the GSOsvm has turn out to be best kernel function with local features to a global representative for any type of dataset. As a final point, experiments are done for dataset with different groups that are formed to show the superior value based on its accuracy on prediction of this kind of model which proves the best validation. Though many research readings suggest the usage of Radial basis (RBF) kernels and polynomial kernels for the conventional techniques, it was found that the results produced by these models have unreliable values, because of the sensitive in the data. The new kernel GSOsvm has the good reliable results both for real and simulated data values precisely when the data contains extreme observations that violated assumptions.

Short-term forecasting of COVID-19 using support vector regression: An application using Zimbabwean data

This study aims to show that including pairwise hierarchical interactions of covariates and combining forecasts from individual models improves prediction accuracy. The least absolute shrinkage and selection operator via hierarchical pairwise interaction is used in selecting variables that are not correlated and with the greatest predictive power in single forecast models (Gradient boosting method [GBM], Generalized additive models [GAMs], Support vector regression [SVR]) are used in the analysis. The best model was selected based on the mean absolute error (MAE), the best key performance indicator for skewed data. Forecasts from the 5 models were combined using linear quantile regression averaging (LQRA). Box and Whiskers plots are used to diagnose the overall performance of fitted models. Single forecast models (GBM, GAMs, and SVRs) show that including pairwise interactions improves forecast accuracy. The SVR model with interactions based on the radial basis kernel function is the best from single forecast models with the lowest MAE. Combining point forecasts from all the single forecast models using the LQRA approach further reduces the MAE. However, based on the Box and Whiskers plot, the SVR model with pairwise interactions has the smallest range. Based on the key performance indicators, combining predictions from several individual models improves forecast accuracy. However, overall, the SVM with pairwise hierarchical interactions outperforms all the other models.

Raman spectroscopy combined with a support vector machine algorithm as a diagnostic technique for primary Sjögren’s syndrome

The aim of this study was to explore the feasibility of Raman spectroscopy combined with computer algorithms in the diagnosis of primary Sjögren syndrome (pSS). In this study, Raman spectra of 60 serum samples were acquired from 30 patients with pSS and 30 healthy controls (HCs). The means and standard deviations of the raw spectra of patients with pSS and HCs were calculated. Spectral features were assigned based on the literature. Principal component analysis (PCA) was used to extract the spectral features. Then, a particle swarm optimization (PSO)-support vector machine (SVM) was selected as the method of parameter optimization to rapidly classify patients with pSS and HCs. In this study, the SVM algorithm was used as the classification model, and the radial basis kernel function was selected as the kernel function. In addition, the PSO algorithm was used to establish a model for the parameter optimization method. The training set and test set were randomly divided at a ratio of 7:3. After PCA dimension reduction, the specificity, sensitivity and accuracy of the PSO-SVM model were obtained, and the results were 88.89%, 100% and 94.44%, respectively. This study showed that the combination of Raman spectroscopy and a support vector machine algorithm could be used as an effective pSS diagnosis method with broad application value.

Prediction Method for Sugarcane Syrup Brix Based on Improved Support Vector Regression

The brix of syrup is an important parameter in sugar production. To accurately measure syrup brix, a novel measurement method based on support vector regression (SVR) is presented. With the resonant frequency and quality factor as inputs and syrup brix as the output, a mathematical model of the relationship between the resonant frequency, quality factor, and syrup brix is established. Simultaneously, the particle swarm optimization (PSO) algorithm is used to optimize the penalty coefficient and radial basis kernel function of SVR to improve the performance of the model. The calculation model is trained and tested using the collected experimental data. The results show that the mean absolute error, mean absolute percentage error, and root mean square error of the syrup brix calculation model based on the improved SVR model can reach 0.74 °Bx, 2.24%, and 0.90 °Bx, respectively, while the determination coefficient can reach 0.9985. The simulation of the online measurement of syrup brix in the actual production process proves the excellent prediction performance of the syrup brix calculation model based on the improved PSO–SVR model, which can thus be used to predict the syrup brix.

Rapid detection of serological biomarkers in gallbladder carcinoma using fourier transform infrared spectroscopy combined with machine learning

Gallbladder cancer (GBC) is the most common malignant tumour of the biliary tract. GBC is difficult to diagnose and treat at an early stage because of the lack of effective serum markers and typical symptoms, resulting in low survival rates. This study aimed to investigate the applicability of dried serum Fourier-transform infrared (FTIR) spectroscopy combined with machine learning algorithms to correctly differentiate patients with GBC from patients with gallbladder disease (GBD), cholangiocarcinoma (CCA), hepatocellular carcinoma (HCC) and healthy individuals. The differentiation between healthy individuals and GBC serum was better using principal component analysis (PCA) and linear discriminant analysis (LDA) for six spectral regions, especially in the protein (1710-1475 cm−1) and combined (1710–1475 + 1354-980 cm−1) region. However, the PCA-LDA model poorly differentiated GBC from GBD, CCA, and HCC in serum spectra. We evaluated the PCA- LDA, PCA-support vector machine (SVM), and radial basis kernel function support vector machine (RBF-SVM) models for GBC diagnosis and found that the RBF-SVM model performed the best, with 88.24–95% accuracy, 95.83% sensitivity, and 78.38–94.44% specificity in the 1710–1475 + 1354-980 cm−1 region. This study demonstrated that serum FTIR spectroscopy combined with the RBF-SVM algorithm has great clinical potential for GBC screening.

A fault diagnosis method for building electrical systems based on the combination of variational modal decomposition and new mutual dimensionless

The fault diagnosis of building electrical systems are of great significance to the safe and stable operation of modern intelligent buildings. In this paper, it has many problems, such as various fault types, inconspicuous fault characteristics, uncertainty of fault type and mode, irregularity, unstable signal, large gap between fault data classes, small gap between classes and nonlinearity, etc. A method of building electrical system fault diagnosis based on the combination of variational mode decomposition and mutual dimensionless indictor (VMD-MDI) and quantum genetic algorithm-support vector machine (QGA-SVM) is proposed. Firstly, the method decomposes the original signal through variational modal decomposition to obtain the optimal number of Intrinsic Mode Function(IMF) containing fault feature information. Secondly, extracts the mutual dimensionless indicator for each IMF. Thirdly, the optimal penalty coefficient C of the support vector machine and the parameter gamma (gamma) in the radial basis kernel function are selected by the quantum genetic algorithm. Finally, SVM optimized by the QGA is used to identify and classify the faults. By applying the proposed method to the experimental platform data of building electrical system, and compared with the traditional feature extraction method Empirical Mode Decomposition (EMD), Singular Value Decomposition(SVD), Local Mean Decomposition(LMD). And compared with traditional SVM, Genetic Algorithm optimized Support Vector Machine (GA-SVM), One-Dimensional Convolutional Neural Network (1DCNN) for fault classification methods. The experimental results show that the method has better effect and higher accuracy in fault diagnosis and classification of building electrical system. Its average test accuracy can reach 91.67%.

Support Vector Regression Models of Stormwater Quality for a Mixed Urban Land Use

The present study is an attempt to model the stormwater quality of a stream located in Pune, India. The city is split up into twenty-three basins (named A to W) by the Pune Municipal Corporation. The selected stream lies in the haphazardly expanded peri-urban G basin. The G basin has constructed stormwater drains which open up in this selected open stream. The runoff over the regions picks up the non-point source pollutants which are also added to the selected stream. The study becomes more complex as the stream is misused to dump trash materials, garbage and roadside litter, which adds to the stormwater pollution. Experimental investigations include eleven distinct locations on a naturally occurring stream in the G basin. Stormwater samples were collected for twenty-two storm events, for the monsoon season over four years from 2018–2021, during and after rainfall. The physicochemical characteristics were analyzed for twelve water quality parameters, including pH, Conductivity, Turbidity, Total solids (TS), Total Suspended Solids (TSS), Total Dissolved Solids (TDS), Bio-chemical Oxygen Demand (BOD5), Chemical Oxygen Demand (COD), Dissolved Oxygen (DO), Phosphate, Ammonia and Nitrate. The Water Quality Index (WQI) ranged from 46.9 to 153.9 and from 41.20 to 87.70 for samples collected during and immediately after the rainfall, respectively. Principal Component Analysis was used to extract the most significant stormwater quality parameters. To understand the non-linear complex relationship of rainfall characteristics with significant stormwater pollutant parameters, a Support Vector Regression (SVR) model with Radial Basis Kernel Function (RBF) was developed. The Support Vector Machine is a powerful supervised algorithm that works best on smaller datasets but on complex ones with the help of kernel tricks. The accuracy of the model was evaluated based on normalized root-mean-square error (NRMSE), coefficient of determination (R2) and the ratio of performance to the interquartile range (RPIQ). The SVR model depicted the best performance for parameter TS with NRMSE (0.17), R2 (0.82) and RPIQ (2.91). The unit increase or decrease in the coefficients of rainfall characteristics displays the weighted deviation in the values of pollutant parameters. Non-linear Support Vector Regression models confirmed that both antecedent dry days and rainfall are correlated with significant stormwater quality parameters. The conclusions drawn can provide effective information to decision-makers to employ an appropriate treatment train approach of varied source control measures (SCM) to be proposed to treat and mitigate runoff in an open stream. This holistic approach serves the stakeholder’s objectives to manage stormwater efficiently. The research can be further extended by selecting a multi-criteria decision-making tool to adopt the best SCM and its multiple potential combinations.

Automated recognition of asymmetric gait and fatigue gait using ground reaction force data.

Introduction: The purpose of this study was to evaluate the effect of running-induced fatigue on the characteristic asymmetry of running gait and to identify non-linear differences in bilateral lower limbs and fatigued gait by building a machine learning model. Methods: Data on bilateral lower limb three-dimensional ground reaction forces were collected from 14 male amateur runners before and after a running-induced fatigue experiment. The symmetry function (SF) was used to assess the degree of symmetry of running gait. Statistical parameter mapping (Paired sample T-test) algorithm was used to examine bilateral lower limb differences and asymmetry changes pre- and post-fatigue of time series data. The support vector ma-chine (SVM) algorithm was used to recognize the gait characteristics of both lower limbs before and after fatigue and to build the optimal algorithm model by setting different kernel functions. Results: The results showed that the ground reaction forces were asymmetrical (SF > 0.5) both pre-and post-fatigue and mainly concentrated in the medial-lateral direction. The asymmetry of the medial-lateral direction increased significantly after fatigue (p < 0.05). In addition, we concluded that the polynomial kernel function could make the SVM model the most accurate in classifying left and right gait features (accuracy of 85.3%, 82.4%, and 82.4% in medial-lateral, anterior-posterior and vertical directions, respectively). Gaussian radial basis kernel function was the optimal kernel function of the SVM algorithm model for fatigue gait recognition in the medial-lateral and vertical directions (accuracy of 54.2% and 62.5%, respectively). Moreover, polynomial was the optimal kernel function of the anterior-posterior di-rection (accuracy = 54.2%). Discussion: We proved in this study that the SVM algorithm model depicted good performance in identifying asymmetric and fatigue gaits. These findings can provide implications for running injury prevention, movement monitoring, and gait assessment.

Efficient Selection of Gaussian Kernel SVM Parameters for Imbalanced Data

For medical data mining, the development of a class prediction model has been widely used to deal with various kinds of data classification problems. Classification models especially for high-dimensional gene expression datasets have attracted many researchers in order to identify marker genes for distinguishing any type of cancer cells from their corresponding normal cells. However, skewed class distributions often occur in the medical datasets in which at least one of the classes has a relatively small number of observations. A classifier induced by such an imbalanced dataset typically has a high accuracy for the majority class and poor prediction for the minority class. In this study, we focus on an SVM classifier with a Gaussian radial basis kernel for a binary classification problem. In order to take advantage of an SVM and to achieve the best generalization ability for improving the classification performance, we will address two important problems: the class imbalance and parameter selection during SVM parameter optimization. First of all, we proposed a novel adjustment method called b-SVM, for adjusting the cutoff threshold of the SVM. Second, we proposed a fast and simple approach, called the Min-max gamma selection, to optimize the model parameters of SVMs without carrying out an extensive k-fold cross validation. An extensive comparison with a standard SVM and well-known existing methods are carried out to evaluate the performance of our proposed algorithms using simulated and real datasets. The experimental results show that our proposed algorithms outperform the over-sampling techniques and existing SVM-based solutions. This study also shows that the proposed Min-max gamma selection is at least 10 times faster than the cross-validation selection based on the average running time on six real datasets.

Efficient reinforcement learning with least-squares soft Bellman residual for robotic grasping

Grasping control of intelligent robots has to deal with the difficulties of model uncertainties and nonlinearities. In this paper, we propose the Kernel-based Least-Squares Soft Bellman residual Actor–Critic (KLSAC) algorithm for robotic grasping. In the proposed approach, a novel linear temporal-difference learning algorithm using the least-squares soft Bellman residual (LS2BR) method is designed for policy evaluation. In addition, KLSAC adopts a sparse-kernel feature representation method based on approximate linear dependency (ALD) analysis to construct features for continuous state–action space. Compared with typical deep reinforcement learning algorithms, KLSAC has two main advantages: firstly, the critic module has the capacity for rapid convergence by computing the fixed point of the linear soft Bellman equation via the least-squares optimization method. Secondly, the kernel-based features construction approach only requires predefining the basic kernel function and can improve the generalization ability of KLSAC. The simulation studies on robotic grasping control were conducted in the V-REP simulator. The results demonstrate that compared with other typical RL algorithms (e.g., SAC and BMPO), the proposed KLSAC algorithm can achieve better performance in terms of sample efficiency and asymptotic convergence property. Furthermore, experimental results on a real UR5 robot validated that KLSAC performed well in the real world.

Roughness detection method based on image multi-features

Roughness was one of the most visual manifestations of the surface quality of metal parts. It affected the performance and life of the parts. Accurate and efficient roughness grade detection technology was of great significance to smart manufacturing. Traditional machine shops often used roughness comparison sample blocks and stylus profilers to check roughness. However, there were disadvantages such as slow detection speed and high influence by human factors. As a non-destructive testing technique, optical imaging gad already demonstrated to be an effective roughness inspection method. In this paper, a roughness detection approach based on image multi-features was proposed, using part surface images as the research object. First, gray level co-occurrence matrix (GLCM), Gabor transform, and local binary patterns (LBP) were used for the extraction of image texture features. After using principal components analysis to reduce the dimensionality of texture features, multiple texture features were concatenated to form a multi-feature vector. Finally, the multi-feature vectors were input into the Gaussian radial basis kernel support vector machine to classify the part surface images and thus completed the detection of roughness grade.

Automated Detection of Seizure Types from the Higher-Order Moments of Maximal Overlap Wavelet Distribution.

In this work, an attempt has been made to develop an automated system for detecting electroclinical seizures such as tonic-clonic seizures, complex partial seizures, and electrographic seizures (EGSZ) using higher-order moments of scalp electroencephalography (EEG). The scalp EEGs of the publicly available Temple University database are utilized in this study. The higher-order moments, namely skewness and kurtosis, are extracted from the temporal, spectral, and maximal overlap wavelet distributions of EEG. The features are computed from overlapping and non-overlapping moving windowing functions. The results show that the wavelet and spectral skewness of EEG is higher in EGSZ than in other types. All the extracted features are found to have significant differences (p < 0.05), except for temporal kurtosis and skewness. A support vector machine with a radial basis kernel designed using maximal overlap wavelet skewness yields a maximum accuracy of 87%. In order to improve the performance, the Bayesian optimization technique is utilized to determine the suitable kernel parameters. The optimized model achieves the highest accuracy of 96% and an MCC of 91% in three-class classification. The study is found to be promising, and it could facilitate the rapid identification process of life-threatening seizures.

Modeling of scour depth and length of a diversion channel flow system with soft computing techniques

Abstract This study employed soft computing techniques, namely, support vector machine (SVM) and Gaussian process regression (GPR) techniques, to predict the properties of a scour hole [depth (ds) and length (Ls)] in a diversion channel flow system. The study considered different geometries of diversion channels (angles and bed widths) and different hydraulic conditions. Four kernel function models for each technique (polynomial kernel function, normalized polynomial kernel function, radial basis kernel, and the Pearson VII function kernel) were evaluated in this investigation. Root mean square error (RMSE) values are 8.3949 for training datasets and 11.6922 for testing datasets, confirming that the normalized polynomial kernel function-based GP outperformed other models in predicting Ls. Regarding predicting ds, the polynomial kernel function-based SVM outperforms other models, recording RMSE of 0.5175 for training datasets and 0.6019 for testing datasets. The sensitivity investigation of input parameters shows that the diversion angle had a major influence in predicting Ls and ds.

A Comparative Analysis of Machine Learning Models: A Case Study in Predicting Chronic Kidney Disease

In the modern world, chronic kidney disease is one of the most severe diseases that negatively affects human life. It is becoming a growing problem in both developed and underdeveloped countries. An accurate and timely diagnosis of chronic kidney disease is vital in preventing and treating kidney failure. The diagnosis of chronic kidney disease through history has been considered unreliable in many respects. To classify healthy people and people with chronic kidney disease, non-invasive methods like machine learning models are reliable and efficient. In our current work, we predict chronic kidney disease using different machine learning models, including logistic, probit, random forest, decision tree, k-nearest neighbor, and support vector machine with four kernel functions (linear, Laplacian, Bessel, and radial basis kernels). The dataset is a record taken as a case–control study containing chronic kidney disease patients from district Buner, Khyber Pakhtunkhwa, Pakistan. To compare the models in terms of classification and accuracy, we calculated different performance measures, including accuracy, Brier score, sensitivity, Youdent, specificity, and F1 score. The Diebold and Mariano test of comparable prediction accuracy was also conducted to determine whether there is a substantial difference in the accuracy measures of different predictive models. As confirmed by the results, the support vector machine with the Laplace kernel function outperforms all other models, while the random forest is competitive.

Covid-19'da Pozitif Vaka Oranı Tahmini İçin Aşırı Öğrenme Makinesi Algoritmaları: Karşılaştırmalı Bir Çalışma

Various pandemics have been recorded in world history until today. The Covid-19 outbreak, which emerged at the end of 2019, has recently been a hot topic in the literature. In this work, extreme learning algorithms are presented as a comparative study for predicting the positive rate for the countries: India, Turkey, Italy, USA and UK. The features to be used in the learning phase are determined with the F-test feature selection method. For each extreme learning approach, results are obtained for each country with the root mean square error evaluation criteria. Accordingly, the radial basis kernel function produces the best estimation results, while the linear kernel function has the highest RMSE. Accordingly, the lowest RMSE value has been obtained for India as 4.17E-03 with the radial basis kernel function based ELM. Also, since Turkey's data contains too many outliers, it has the highest RMSE value (0.015 - 0.035) in linear kernel method among the countries.

Optimization of Support Vector Machine Algorithm Using Stunting Data Classification

Several studies from Indonesia reveal that malnutrition and stunting are still severe concerns to be addressed in the future. The complexity of the problem of stunting or nutritional status requires the responsibility of all parties, including science and technology. The issue of monitoring and data collection related to stunting or the nutritional status of children in Indonesia, especially Medan City, North Sumatra Province, is an essential factor in determining the calculations carried out by each Community Health Center with many attributes. Currently, the Support Vector Machine method is a solution to increase government intervention's effectiveness in classifying malnutrition and stunting. However, the Support Vector Machine algorithm still needs to improve, namely the difficulty of selecting the right and optimal features for the attribute weights, causing a low prediction accuracy. Therefore, researchers aim to optimize the Support Vector Machine Algorithm with Particle Swarm Optimization using Linear, Polynomial, Sigmoid, and Radial Basic Function kernels. The results were obtained from research utilizing nutritional status data, that performance in improving the Support Vector Machine algorithm based on Particle Swarm Optimization using four kernel tests, namely Linear, Polynomial, Sigmoid, and Radial Basic Function obtained different results, not all kernels in this study can improve accuracy well. The best performance is using the Radial Basic Function kernel with an Accuracy value of 78%, Precision of 89%, Recall of 66%, and F1-Score of 72%, so it is feasible for accurate information regarding the classification of nutritional status.

Feature Map Compression for Video Coding for Machines Based on Receptive Block Based Principal Component Analysis

This paper presents a method to effectively compress the intermediate layer feature map of a convolutional neural network for the potential structures of Video Coding for Machines, which is an emerging technology for future machine consumption applications. Notably, most extant studies compress a single feature map and hence cannot entirely consider both global and local information within the feature map. This limits performance maintenance during machine consumption tasks that analyze objects with various sizes in images/videos. To address this problem, a multiscale feature map compression method is proposed that consists of two major processes: receptive block based principal component analysis (RPCA) and uniform integer quantization. The RPCA derives the complete basis kernels of a feature map by selecting a set of major basis kernels that can represent a sufficient percentage of global or local information according to the variable-size receptive blocks of each feature map. After transforming each feature map using the set of major basis kernels, a uniform integer quantizer converts the 32-bit floating-point values of the set of major basis kernels, corresponding RPCA coefficients, and a mean vector to five-bit integer representation values. Experiment results reveal that the proposed method reduces the amount of feature maps by 99.30% with a loss of 8.30% in the average precision ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">AP</i> ) on the OpenImageV6 dataset and 0.77% in <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">AP_M</i> and 0.47% in <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">AP_L</i> on the MS COCO 2017 validation set while outperforming previous PCA-based feature map compression methods even at higher compression rates.