Search

Data augmentation strategies have always been important in machine learning techniques and play a unique role in model performance optimization processes. Therefore, in recent years, these techniques have become popular in the artificial intelligence field. In this paper, a new data augmentation strategy is proposed based on the interpretation algorithm of deep convolutional neural networks, i.e., constructing new training samples by deeply exploiting key features extracted from interpretable networks to achieve sample augmentation. Thus, a novel supervised data augmentation approach known as Supervised Data Augmentation–Key Feature Extraction (SDA-KFE) was proposed. By introducing the Neural Network Interpreter-Segmentation Recognition and Interpretation (NNI-SRI) algorithm, an augmentation strategy is proposed that can balance the high accuracy and high robustness of the final model while ensuring a large amount of data augmentation. The advantages of the SDA-KFE algorithm are mainly reflected in the following aspects. First, it is easy to implement. This algorithm is implemented based on the lightweight NNI-SRI algorithm, which lays the foundation for the implementation of SDA-KFE so that it can be easily implemented on convolutional neural networks. Second, this model, which is widely applicable, can be applied to almost any deep convolutional network. Through research and experiments on this proposed algorithm, SDA-KFE can be applied in graphical image binary classification and multiclassification models. Third, SDA-KFE can rapidly construct data samples with diverse variations. Under the premise of determining the classification labels of the generated samples, the distribution of the feature unit composition of the samples can be controlled. Compared with traditional data augmentation methods, SDA-KFE can control the direction of the model performance, i.e., the balance between the pursuit of high accuracy and robust performance of the model. Therefore, the novel supervised augmentation approach proposed in this paper is relevant for optimizing deep convolutional neural networks, solving model overfitting, augmenting data types, etc. The data augmentation algorithm proposed in this paper can be regarded as a useful supplement to traditional data augmentation methods, such as horizontal or vertical image flipping, cropping, color transformation, extension and rotation. [ABSTRACT FROM AUTHOR]

Graph-based semi-supervised learning has recently proved to be a powerful paradigm for processing and mining large datasets. The main advantage relies on the fact that these methods can be useful in propagating a small set of known labels to a large set of unlabeled data. However, the lack of labeled data in real applications may affect the semi-supervised learning's performance. This paper introduces a new semi-supervised framework for simultaneous linear feature extraction and label propagation. The proposed method simultaneously estimates a discriminant transformation, as well as the unknown label, by exploiting both labeled and unlabeled data. The learning model's unknowns are to be estimated by integrating two types of graph-based smoothness constraints. The resulting semi-supervised model is expected to learn more discriminative information. Experiments are conducted on nine public image datasets. These experimental results show that the proposed method's performance can be better than that of many state-of-the-art graph-based semi-supervised algorithms. [ABSTRACT FROM AUTHOR]

• A TFADNN method is developed for scene classification of remote sensing images. • The FC layers of pre-trained CNN are replaced by a GAP layer in the first stream. • A multi-scale nonlinear encoding based BoVW model is proposed in the second stream. • A weighted fusion is adopted to integrate two-stream features for classification. • The TFADNN can learn discriminative features from HSR images with arbitrary sizes. Scene classification of high-spatial resolution (HSR) images has a wide range of potential applications in various fields, and it has become a research hotspot in remote sensing community. Recently, deep transfer learning-based methods have attracted tremendous attention due to powerful ability of feature extraction. In this paper, a novel architecture termed two-stream feature aggregation deep neural network (TFADNN) is developed for HSR scene classification. The TFADNN method contains two parallel parts, including the stream of discriminative features and the stream of general features. In the first stream, the fully connected layers of pre-trained CNNs are replaced by a global average pooling layer to remove the limitation on the size of input images. As for the second stream, the multi-scale nonlinear encoding based bag-of-visual-words (MNBoVW) model is proposed to process convolutional features, and the global representations can be obtained. Then, weighted fusion is adopted to integrate two-stream features. As a result, the TFADNN method can learn the discriminative features from HSR images with arbitrary sizes, and the experimental results on two challenging datasets indicate that the TFADNN method achieves satisfactory classification performance compared with some state-of-the-art methods. [ABSTRACT FROM AUTHOR]

• A data-driven model is proposed based on multilevel features fusion to classify the variation trend of silicon content. • Multilevel features are extracted through variable analysis, statistical information, and stacked denoising autoencoder, respectively. • Silicon content trend labels are quantitative descripted by seven different primitives. • Both experimental simulation and industrial application verify the effectiveness and feasibility of the proposed model. The silicon content variation trend, which can reflect the quality of molten iron, provides significant information that can assist in ensuring the smooth operation of a blast furnace. This paper proposes a novel dynamic data-driven model for the online classification of the variation trend for the silicon content. Typically, a dynamic model for the silicon content variation trend primarily relies on process data feature extraction. First, a multilevel features fusion algorithm based on mutual information is developed to extract a rich and robust feature representation. Subsequently, the fused multilevel feature vectors and their corresponding trend labels are fed into a recurrent neural network model to capture the process dynamics and classify the variation trend. An experimental simulation and industrial application verified the effectiveness and feasibility of the proposed method. The classification results can provide guidance to ensure that the quality of molten iron is maintained within the desired range in the ironmaking process. [ABSTRACT FROM AUTHOR]

Symbolic Time Intervals (STIs) describe events having a non-zero time duration, which occur in a wide range of application domains. In this paper, we target the challenge of STIs series classification (STIC), which refers to the categorization of series of STIs. Over the recent years several advancements have been made in STIC, all of which are based on either distance-metrics or feature-based traditional classifiers, mostly relying on hand-engineering of features. Due to the high computational cost of either distance calculation or feature extraction, most methods also have quite little potential to scale. We introduce INSTINCT – a novel deep learning-based framework for STIC, which 1) proposes an almost fully information-preserving transformation of raw STIs series into real matrices, and 2) presents a novel ensemble of deep inception-based convolutional neural networks for their classification. The evaluation is applied to the six real-world STIC benchmark datasets and demonstrates that INSTINCT significantly improves accuracy over seven state-of-the-art methods, as well as over three deep learning-based baselines. In addition, a comprehensive architecture study of INSTINCT is conducted as well as a scalability analysis, reporting an overall time complexity which is linear in each of the main properties of the input STIs series. • Novel deep learning-based framework for Symbolic Time Intervals series classification. • Representation of raw Symbolic Time Intervals series as real matrices. • Inception-based networks ensemble for Symbolic Time Intervals series classification. • Improved classification accuracy over state-of-the-art. • Linear time complexity. [ABSTRACT FROM AUTHOR]

The electrocardiogram (ECG) signal is made up of sequences of three distinct waves, including the P-wave, QRS-complex, and T-wave. These sequences may contain several different varieties of feature representation, which means various appearances represent different attribute knowledge of heart activity. Consequently, delicate analysis of the ECG attribute knowledge is critical, which focuses on empowering computers to understand ECG signals in a way physicians do. Traditional ECG systems usually only focus on the key wave detection or sub-sequences morphological classification, and the complex ECG features make the analysis process difficult to handle both tasks. In this paper, our goal is to propose a novel ECG analysis technique for ECG medical attribute knowledge feature extraction, termed ECG-MAKE method. The ECG-MAKE method mainly addresses two challenging issues—ECG signal key points detection and morphological delineation—by introducing four modules: signal preprocessing, key point detection, adaptive error correction, and attribute knowledge extraction. Signal preprocessing is based on a wavelet multi-resolution algorithm, which extracts both time and frequency domain information simultaneously, and achieves the goal of enhancing raw signals by filtering signals below 1 Hz and above 100 Hz. The key point detection combines mean-value difference and sliding window to detect R-peak. And according to the R-peak, further utilizes the ECG rules and threshold to detect the onset, peak, and offset of different heartbeat waveforms such as the P-waves and T-waves. Then, for standard 12-lead ECG, the adaptive error correction can automatically correct the spatial position of certain leads missed and false detection key points. Finally, the attribute knowledge extraction module is based on the statistical features of key points to obtain the ECG attributes knowledge of clinical diagnosis of ECG. Extensive experiments conducted on the publicly available database QT database (99.57% F1-score) have demonstrated the effectiveness of the proposed ECG-MAKE method. The proposed method has a higher classification performance than current state-of-the-art methods with an F1-score performance of 94.30% on the ZZU-ECG database. • A medical attribute knowledge extraction method is proposed to delineate ECG signal. • A strategy is proposed to correct missed and false detections of ECG knowledge. • A framework is proposed to provide ECG refined annotation for deep learning model. • High classification performance compared to current state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Text categorization systems are designed to classify documents into a fixed number of predefined categories. Bag-of-words is one of the most used approaches to represent a document. However, it generates high-dimensional sparse data matrix with a high feature-to-instance ratio. An aggressive feature selection can alleviate these drawbacks, but such selection degrades the classifier’s performance. In this paper, we propose an approach for text categorization based on Dissimilarity Representation and multiple classifier systems. The proposed system, Combined Dissimilarity Spaces (CoDiS), is composed of multiple classifiers trained on data from different dissimilarity spaces. Each dissimilarity space is a transformation of the original space that reduces the dimensionality, feature-to-instance ratio, and sparseness. Experiments using forty-seven text categorization databases show that CoDiS presents a better performance in comparison to literature systems. [ABSTRACT FROM AUTHOR]

This paper introduces a novel algorithm, called Supervised Aggregated FEature learning or SAFE, which combines both (local) instance level and (global) bag level information in a joint framework to address the multiple instance classification task. In this realm, the collective assumption is used to express the relationship between the instance labels and the bag labels, by means of taking the sum as aggregation rule. The proposed model is formulated within a least squares support vector machine setting, where an unsupervised core model (either kernel PCA or kernel spectral clustering) at the instance level is combined with a classification loss function at the bag level. The corresponding dual problem consists of solving a linear system, and the bag classifier is obtained by aggregating the instance scores. Synthetic experiments suggest that SAFE is advantageous when the instances from both positive and negative bags can be naturally grouped in the same cluster. Moreover, real-life experiments indicate that SAFE is competitive with the best state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Interval-valued observations arise in many real-life situations as either the precise representation of the objective entity or the representation of incomplete knowledge. Thus given p features observed over a sample of objects belonging to one of two possible classes, each observation can be perceived as a non-empty closed and bounded hyperrectangle on R p . The aim of the paper is to suggest a p -dimensional classification method for random intervals when two or more classes are considered, by the generalization of Fisher’s procedure for linear discriminant analysis. The idea consists of finding a directional vector which maximizes the ratio of the dispersion between the classes and within the classes of the observed hyperrectangles. A classification rule for new observations is also provided and some simulations are carried out to compare the behavior of the proposed classification procedure with respect to other methods known from the literature. Finally, the suggested methodology are applied on a real-life situation example. [ABSTRACT FROM AUTHOR]

A binary coding based feature extraction (BCFE) method is proposed in this paper. In the BCFE method, the spectral signature of each pixel of hyperspectral image is partitioned into some equal segments. Then, the weighted mean of spectral bands in each segment is considered as an extracted feature. BCFE uses a new method for calculation of weights. In BCFE, the binary codes of class means are obtained. Then, the information contained in the binary values and the edges of class means is used for calculation of weight in each band. The experimental results on three real hyperspectral images show the better performance of BCFE compared to some popular and state-of-the-art feature extraction methods, from the accuracy and computation time point of views, in a small sample size situation. [ABSTRACT FROM AUTHOR]