Earticle

The requirements of multi-resolution models of feature-based solids, which represent an object at many levels of feature detail, are increasing for engineering purposes, such as analysis, network-based collaborative design, virtual prototyping and manufacturing. To provide multi-resolution models for various applications, it is essential to generate adequate solid models at varying levels of detail (LOD) after feature rearrangement, based on the LOD criteria. However, the non-commutative property of the union and subtraction Boolean operations is a severe obstacle to arbitrary feature rearrangement. To solve this problem we propose history-based Boolean operations that satisfy the commutative law between union and subtraction operations by considering the history of the Boolean operations. Because these operations guarantee the same resulting shape as the original and reasonable shapes at the intermediate LODs for an arbitrary rearrangement of its features, various LOD criteria can be applied for multi-resolution modeling in different applications.

We propose a feature-based multi-resolution representation of B-rep solid models using history-based Boolean operations based on the merge-and-select algorithm. Because union and subtraction are commutative in the history-based Boolean operations, the integrity of the models at various levels of detail (LOD) is guaranteed for the reordered features regardless of whether the features are subtractive or additive. The multi-resolution solid representation proposed in this paper includes a non-manifold topological merged-set model of all feature primitives as well as a feature-modeling tree reordered consistently with a given LOD criterion. As a result, a B-rep solid model for a given LOD can be provided quickly, because the boundary of the model is evaluated without any geometric calculation and extracted from the merged set by selecting the entities contributing to the LOD model shape.

Various machine-learning models may yield high predictive power for massive time series for time series prediction. However, these models are prone to instability in terms of computational cost because of the high dimensionality of the feature space and nonoptimized hyperparameter settings. Considering the potential risk that model training with a high-dimensional feature set can be time-consuming, we evaluate a feature-importance-based feature selection method to derive a tradeoff between predictive power and computational cost for time series prediction. We used two machine learning techniques for performance evaluation to generate prediction models from a retail sales dataset. First, we ranked the features using impurity- and Local Interpretable Model-agnostic Explanations (LIME) -based feature importance measures in the prediction models. Then, the recursive feature elimination method was applied to eliminate unimportant features sequentially. Consequently, we obtained a subset of features that could lead to reduced model training time while preserving acceptable model performance.

This study developed and implemented a new feature-based map-building model that uses only sparsely sampled sonar data from a fixed ring with 16 sonar sensors. It introduces two kinds of data filter approaches to overcome challenges associated with sonar sensors, such as a wide beam aperture and the specular reflection effect. The first approach is a footprint-association (FPA) model, which associates two sonar footprints into a hypothesized circle frame in order to determine the feature type, such as a line, a point, or an arc. The FPA model provides information about the trace of centers of hypothesized circles. It extracts features from a cluster composed of more than two independent footprints that originate from the same object. The other approach is a feature-association (FTA) model, which associates a new sonar footprint into extracted features to update the feature. Both proposed methods were tested in a home-like environment using a mobile robot.

The success of iris recognition depends mainly on two factors: image acquisition and an iris recognition algorithm. In this study, we present a system that considers both factors and focuses on the latter. The proposed algorithm aims to find out the most efficient wavelet family and its coefficients for encoding the iris template of the experiment samples. The algorithm implemented in software performs segmentation, normalization, feature encoding, data storage, and matching. By using the Haar and Biorthogonal wavelet families at various levels feature encoding is performed by decomposing the normalized iris image. The vertical coefficient is encoded into the iris template and is stored in the database. The performance of the system is evaluated by using the number of degrees of freedom, False Reject Rate (FRR), False Accept Rate (FAR), and Equal Error Rate (EER) and the metrics show that the proposed algorithm can be employed for an iris recognition system.

In order to reconstruct a full 3D human model in reverse engineering (RE), a 3D scanner needs to be placed arbitrarily around the target model to capture all part of the scanned surface. Then, acquired multiple scans must be registered and merged since each scanned data set taken from different position is just given in its own local co-ordinate system. The goal of the registration is to create a single model by aligning all individual scans. It usually consists of two sub-steps: rough and fine registration. The fine registration process can only be performed after an initial position is approximated through the rough registration. Hence an automated rough registration process is crucial to realize a completely automatic RE system. In this paper an automated rough registration method for aligning multiple scans of complex human face is presented. The proposed method automatically aligns the meshes of different scans with the information of features that are extracted from the estimated principal curvatures of triangular meshes of the human face. Then the roughly aligned scanned data sets are further precisely enhanced with a fine registration step with the recently popular Iterative Closest Point (ICP) algorithm. Some typical examples are presented and discussed to validate the proposed system.

Convolutional neural network-based deep learning technology is the most commonly used in image identification, but it requires large-scale data for training. Therefore, application in specific fields in which data acquisition is limited, such as in the military, may be challenging. In particular, the identification of ground weapon systems is a very important mission, and high identification accuracy is required. Accordingly, various studies have been conducted to achieve high performance using small-scale data. Among them, the ensemble method, which achieves excellent performance through the prediction average of the pre-trained models, is the most representative method; however, it requires considerable time and effort to find the optimal combination of ensemble models. In addition, there is a performance limitation in the prediction results obtained by using an ensemble method. Furthermore, it is difficult to obtain the ensemble effect using models with imbalanced classification accuracies. In this paper, we propose a transfer learning-based feature fusion technique for heterogeneous models that extracts and fuses features of pre-trained heterogeneous models and finally, fine-tunes hyperparameters of the fully connected layer to improve the classification accuracy. The experimental results of this study indicate that it is possible to overcome the limitations of the existing ensemble methods by improving the classification accuracy through feature fusion between heterogeneous models based on transfer learning.

According to the fast growth of information on the Internet, it is becoming increasingly difficult to find and organize useful information. To reduce information overload, it needs to exploit automatic text classification for handling enormous documents. Support Vector Machine (SVM) is a model that is calculated as a weighted sum of kernel function outputs. This paper describes a document classifier for web documents in the fields of Information Technology and uses SVM to learn a model, which is constructed from the training sets and its representative terms. The basic idea is to exploit the representative terms meaning distribution in coherent thematic texts of each category by simple statistics methods. Vector-space model is applied to represent documents in the categories by using feature selection scheme based on TFiDF. We apply a category factor which represents effects in category of any term to the feature selection. Experiments show the results of categorization and the correlation of vector length.

Identifying informative features in bioinformatics is challenging due to their small proportion within large datasets. We propose a scalable and interpretable feature selection framework for cancer RNA-seq by transforming non-image bio-data into 2D formats and applying convolutional neural networks (CNNs) with transfer learning for efficient classification. Explainable artificial intelligence (XAI) techniques identify and prioritize important features, while principal component analysis (PCA) determines the optimal number of selected features, ensuring transparency and reliability. Comparative analysis of CNN and XAI highlights the effectiveness of our approach, providing a robust framework for high-dimensional genomic data analysis with applications in cancer diagnosis and prognosis.

An innovative tone modeling framework based on deep neural networks in tone recognition was proposed in this paper. In the framework, both the prosodic features and the articulatory features were firstly extracted as the raw input data. Then, a 5-layer-deep deep belief network was presented to obtain high-level tone features. Finally, support vector machine was trained to recognize tones. The 863-data corpus had been applied in experiments, and the results show that the proposed method helped improve the recognition accuracy significantly for all tone patterns. Meanwhile, the average tone recognition rate reached 83.03%, which is 8.61% higher than that of the original method.

In this paper, a novel method is proposed to build an ensemble of classifiers by using a feature selection schema. The feature selection schema identifies the best feature sets that affect the arrhythmia classification. Firstly, a number of feature subsets are extracted by applying the feature selection schema to the original dataset. Then classification models are built by using the each feature subset. Finally, we combine the classification models by adopting a voting approach to form a classification ensemble. The voting approach in our method involves both classification error rate and feature selection rate to calculate the score of the each classifier in the ensemble. In our method, the feature selection rate depends on the extracting order of the feature subsets. In the experiment, we applied our method to arrhythmia dataset and generated three top disjointed feature sets. We then built three classifiers based on the top-three feature subsets and formed the classifier ensemble by using the voting approach. Our method can improve the classification accuracy in high dimensional dataset. The performance of each classifier and the performance of their ensemble were higher than the performance of the classifier that was based on whole feature space of the dataset. The classification performance was improved and a more stable classification model could be constructed with the proposed approach.

Automatic segmentation of foreground text from the background in degraded document images is very much essential for the smooth reading of the document content and recognition tasks by machine. In this paper, we present a novel approach to the binarization of degraded document images. The proposed method uses a new local contrast feature extracted based on the stroke width of text. First, a pre-processing method is carried out for noise removal. Text boundary detection is then performed on the image constructed from the contrast feature. Then local estimation follows to extract text from the background. Finally, a refinement procedure is applied to the binarized image as a post-processing step to improve the quality of the final results. Experiments and comparisons of extracting text from degraded handwriting and machine-printed document image against some well-known binarization algorithms demonstrate the effectiveness of the proposed method.

Deep convolution neural networks are the recent algorithms used for robotic vision. However, the complex crop?weed vegetation and the background interferences required a robust feature representation. Therefore, we proposed a Dual-branch Deep neural network for the semantic segmentation of crops and weeds. The branches utilized distinct feature extraction algorithms that extract essential semantic cues, and a decoder combined these features to improve the global contextual information. Finally, the hybrid feature selection module(HSFM) utilized the decoder features to complement one another. Experimental results show the proposed method obtained mean intersection of union scores of 0.8613 and 0.9099 on CWFID and BoniRob datasets, respectively.

The paper proposes a novel gait recognition algorithm based on feature fusion of gait energy image (GEI) dynamic region and Gabor, which consists of four steps. First, the gait contour images are extracted through the object detection, binarization and morphological process. Secondly, features of GEI at different angles and Gabor features with multiple orientations are extracted from the dynamic part of GEI, respectively. Then averaging method is adopted to fuse features of GEI dynamic region with features of Gabor wavelets on feature layer and the feature space dimension is reduced by an improved Kernel Principal Component Analysis (KPCA). Finally, the vectors of feature fusion are input into the support vector machine (SVM) based on multi classification to realize the classification and recognition of gait. The primary contributions of the paper are: a novel gait recognition algorithm based on based on feature fusion of GEI and Gabor is proposed; an improved KPCA method is used to reduce the feature matrix dimension; a SVM is employed to identify the gait sequences. The experimental results suggest that the proposed algorithm yields over 90% of correct classification rate, which testify that the method can identify better different human gait and get better recognized effect than other existing algorithms.

This paper presents a novel convolutional neural network based multi-feature fusion learning method for non-rigid 3D model retrieval, which can investigate the useful discriminative information of the heat kernel signature (HKS) descriptor and the wave kernel signature (WKS) descriptor. At first, we compute the 2D shape distributions of the two kinds of descriptors to represent the 3D model and use them as the input to the networks. Then we construct two convolutional neural networks for the HKS distribution and the WKS distribution separately, and use the multi-feature fusion layer to connect them. The fusion layer not only can exploit more discriminative characteristics of the two descriptors, but also can complement the correlated information between the two kinds of descriptors. Furthermore, to further improve the performance of the description ability, the cross-connected layer is built to combine the low-level features with high-level features. Extensive experiments have validated the effectiveness of the designed multi-feature fusion learning method.

Effective identification of wireless channel in different scenarios or regions can solve the problems of multipath interference in process of wireless communication. In this paper, different characteristics of wireless channel are extracted based on the arrival time and received signal strength, such as the number of multipath, time delay and delay spread, to establish the feature vector set of wireless channel which is used to train backpropagation (BP) neural network to identify different wireless channels. Experimental results show that the proposed algorithm can accurately identify different wireless channels, and the accuracy can reach 97.59%.

Current methods for semantic segmentation of remote-sensing images, especially for irregular and small targets, often result in low precision and incomplete feature extraction. To address this issue, an improved semantic segmentation method was developed utilizing DeepLabv3+. First, DeepLabv3+ is combined with the proposed feature fusion module to make full use of the complementary information of low- and high-level features. Second, the channel attention module helps extract effective features while suppressing irrelevant features, thereby enabling the extraction of more meaningful global information from high-level features. Finally, rich spatial information is selected using guided spatial attention, which improves the accuracy of edge segmentation of target objects. The results of the comparison show that the mean F1 score (MF1) and overall accuracy (OA) of the proposed method on the ISPRS Potsdam dataset are 89.81% and 88.45%, respectively. The MF1 of the proposed method is 89.90% and the OA is 89.14% for the UAVid dataset, which are higher than those of the other comparison algorithms. The proposed method exhibits superior semantic segmentation capabilities for remote-sensing images.

Coronary heart disease (CHD) is a comorbidity of COVID-19; therefore, routine early diagnosis is crucial. A large number of examination attributes in the context of diagnosing CHD is a distinct obstacle during the pandemic when the number of health service users is significant. The development of a precise machine learning model for diagnosis with a minimum number of examination attributes can allow examinations and healthcare actions to be undertaken quickly. This study proposes a CHD diagnosis model based on feature selection, data balancing, and ensemble-based classification methods. In the feature selection stage, a hybrid SVM-GA combined with fast correlation-based filter (FCBF) is used. The proposed system achieved an accuracy of 94.60% and area under the curve (AUC) of 97.5% when tested on the z-Alizadeh Sani dataset and used only 8 of 54 inspection attributes. In terms of performance, the proposed model can be placed in the very good category.

This paper presents a feature selection technique based on distributional differences for efficient machine learning. Initial training data consists of data including many features and a target value. We classified them into positive and negative data based on the target value. We then divided the range of the feature values into 10 intervals and calculated the distribution of the intervals in each positive and negative data. Then, we selected the features and the intervals of the features for which the distributional differences are over a certain threshold. Using the selected intervals and features, we could obtain the reduced training data. In the experiments, we will show that the reduced training data can reduce the training time of the neural network by about 40%, and we can obtain more profit on simulated stock trading using the trained functions as well.