Self-supervised learning based on discriminative nonlinear features for image classification

For learning-based tasks such as image classification, the feature dimension is usually very high. The learning is afflicted by the curse of dimensionality as the search space grows exponentially with the dimension. Discriminant-EM (DEM) proposed a framework by applying self-supervised learning in a discriminating subspace. This paper extends the linear DEM to a nonlinear kernel algorithm, Kernel DEM (KDEM), and evaluates KDEM extensively on benchmark image databases and synthetic data. Various comparisons with other state-of-the-art learning techniques are investigated for several tasks of image classification. Extensive results show the effectiveness of our approach.     

加权FCM原型空间特征提取的高光谱图像分类

在原型空间特征提取方法的基础上提出一种基于加权原型空间特征提取的方法用于高光谱图像数据分类。通过加权模糊[C]均值算法对每个特征施加不同的权重,从而保证提取后的特征含有较高的信息量。实验结果表明,与PSFE相比,w-PSFE对数据集大小的稳定性更高,同时在提取少量的特征用于高光谱图像数据分类时分类精度更高。     

应用级联多分类器的高光谱图像分类

目的 高光谱分类任务中,由于波段数量较多,图像中存在包含噪声以及各类地物样本分布不均匀等问题,导致分类精度与训练效率不能平衡,在小样本上分类精度低。因此,提出一种基于级联多分类器的高光谱图像分类方法。方法 首先采用主成分分析方法将高度相关的高维特征合成无关的低维特征,以加快Gabor滤波器提取纹理特征的速度;然后使用Gabor滤波器提取图像在各个尺寸、方向上的纹理信息,每一个滤波器会生成一张特征图,在特征图中以待分类样本为中心取一个d×d的邻域,计算该邻域内数据的均值和方差来作为待分类样本的空间信息,再将空间信息和光谱信息融合,以降低光线与噪声的影响;最后将谱—空联合特征输入级联多分类器中,得到预测样本关于类别的概率分布的平均值。结果 实验采用Indian Pines、Pavia University和Salinas 3个数据集,与经典算法如支持向量机和卷积神经网络进行比较,并利用总体分类精度、平均分类精度和Kappa系数作为评价标准进行分析。本文方法总体分类精度在3个数据集上分别达到97.24%、99.57%和99.46%,相对于基于径向基神经网络（RBF）核函数的支持向量机方法提高了13.2%、4.8%和5.68%,相对于加入谱—空联合特征的RBF-SVM （radial basis function-support vector machine）方法提高了2.18%、0.36%和0.83%,相对于卷积神经网络方法提高了3.27%、3.2%和0.3%;Kappa系数分别是0.968 6、0.994 3和0.995 6,亦有提高。结论 实验结果表明,本文方法应用于高光谱图像分类具有较优的分类效果,训练效率较高,无需依赖GPU,而且在小样本上也具有较高的分类精度。     

Hyperspectral remote sensing image classification with information discriminative extreme learning machine

Multimedia Tools and Applications - Hyperspectral remote sensing image classification is important aspect of current research. Extreme learning machine (ELM) has been widely used in the field of...     

基于主动学习的高光谱图像分类方法

高光谱图像监督分类中,为了避免休斯效应需要大量的训练样本,但在实际应用中对样本进行标注成本非常高,因此,得到高质量的训练样本显得十分重要。提出一种基于主动学习的高光谱图像分类方法,通过对区域关注度的统计,有效地结合图像光谱和空间特性,基于主动学习方法获取信息量较大的训练样本,从而较大幅度提高了分类的精确度。实验结果表明,所提算法比传统的随机取样监督分类法和主动学习方法在分类精确度上有较大的优势。     

基于加权K近邻和卷积神经网络的高光谱图像分类

传统的谱空联合分类算法通常定义一个邻域空间作为空间信息,忽略空间中非邻域空间信息,且容易将异类像元也考虑在内。针对于高光谱图像分类问题,提出了一种加权K近邻算法能够自适应地提取空间信息,首先定义光谱和空间坐标组成的特征空间,利用该特征空间寻找目标像元的K个相似像元,并对这些像元根据特征空间进行加权;将加权后的像元按照一定方式组合成三维张量表示最终的谱空联合信息,使用三维卷积神经网络对其进行训练,得到最终分类结果。从实验结果来看,相对于改进前的算法,在总体分类精度上得到了一定的提升,与原始的三维卷积神经网络相比,在收敛速度上也得到大大提升,为高光谱图像的谱空联合分类提供了一种更加实用的方法。     

基于主动深度学习的高光谱影像分类

针对当前高光谱遥感影像分类人工标注样本费时费力,大量未标注样本未得到有效利用以及主要利用光谱信息而忽视空间信息等问题,提出了一种空-谱信息与主动深度学习相结合的高光谱影像分类方法。首先利用主成分分析对原始影像进行降维,在此基础上提取像素的一正方形小邻域作为该像素的空间信息并结合其原始光谱信息得到空谱特征。然后,通过稀疏自编码器得到原始数据的稀疏特征表达,并通过逐层无监督学习稀疏自编码器构建深度神经网络,输出原始数据的深度特征,将其连接到softmax分类器,利用少量标记样本以监督学习的方式完成模型的精调。最后,利用主动学习算法选择最不确定性样本对其进行标注,并加入至训练样本以提高分类器的分类效果。分别对PaviaU影像和PaviaC影像进行分类实验的结果表明,该方法在少量标记样本情况下,相对于传统方法能有效地提高分类精度。     

基于多特征融合半自动岩心图像颗粒提取

基于最大相似度的区域合并算法是一种半自动的图像处理方式,可根据用户提供的交互信息,利用图像特征作为区域相似度进行准确的目标提取。但传统的MSRM算法计算量大,使用单一的图像特征使得分割不够精确。针对这些问题,文章对MSRM算法进行改进,提出一种基于多特征的区域最大相似度图像分割算法,并采用矩阵变换算法来降低计算量。该方法使用超像素图像作为分割基础,首先计算图像相邻区域纹理和颜色特征相似度,并使用矩阵变换算法降低颜色特征矢量维度,然后计算两种特征的权重,最后根据综合后的相似度对图像进行区域合并,得到最终的颗粒提取结果。实验结果表明,该方法可以有效提取颗粒的轮廓,提取的轮廓边缘细节较传统MSRM算法更优,算法执行效率也得到了提高。     

Joint spectral-spatial hyperspectral image classification based on hierarchical subspace switch ensemble learning algorithm

In this paper, a novel spectral-spatial hyperspectral image classification method has been proposed by designing hierarchical subspace switch ensemble learning algorithm. First, the hyperspectral images are processed by fast bilateral filtering to get the spatial features. The spectral features and spatial features are combined to form the initial feature set. Second, Hierarchical instance learning based on iterative means clustering method is designed to obtain hierarchical instance space. Third, random subspace method (RSM) is used for sampling the features and samples, thereby forming multiple sub sample set. After that, semi-supervised learning (S²L) is applied to choose test samples for improving classification performance without touching the class labels. Then, micro noise linear dimension reduction (mNLDR) is used for dimension reduction. Afterwards, ensemble multiple kernels SVM(EMK_SVM) are used for stable classification results. Finally, final classification results are obtained by combining classification results with voting strategy. Experimental results on real hyperspectral scenes demonstrate that the proposed method can effectively improve the classification performance apparently.     

基于空谱融合网络的高光谱图像分类方法

针对高光谱图像分类中提取的空-谱特征表达能力弱及维数较高的问题,提出一种基于空-谱融合网络（SSF-Net）的高光谱图像分类方法。首先,利用双通道卷积神经网络（Two-CNN）同时提取高光谱图像的光谱和空间特征;其次,使用多模态压缩双线性池化（MCB）将所提取的多模态特征向量的外积投射到低维空间,以此产生空-谱联合特征。该特征融合网络,既可以分析光谱特征和空间特征向量中元素之间的复杂关系,同时也避免对光谱和空间向量直接进行外积计算,造成维数过高、计算困难的问题。最终实验表明,与现有基于神经网络的分类方法相比,所提出的高光谱图像分类算法能够获得更高的像元分类精度,表明该网络所提取的空-谱联合向量对高光谱图像具有更强的特征表达能力。     

Three robust features extraction approaches for facial gender classification

This research paper introduces three robust approaches for features extraction for gender classification. The first approach is based on using Discrete Cosine Transform (DCT) and consists of two different methods for calculating features values. The second approach is based on the extraction of texture features using the gray-level cooccurrence matrix (GLCM). The third approach is based on 2D-wavelet transform. The extracted features vectors are classified using SVM. For precise evaluation, the databases used for gender evaluation are based on images from the AT@T, Faces94, UMIST, and color FERET databases. K-fold cross validation is used in training the SVM. The accuracies of gender classification when using one of the two proposed DCT methods for features extraction are 98.6 %, 99.97 %, 99.90 %, and 93.3 % with 2-fold cross validation, and 98.93 %, 100 %, 99.9 %, and 92.18 % with 5-fold cross validation. The accuracies of GLCM texture features approach for facial gender classification are 98.8 %, 99.6 %, 100 %, and 93.11 %, for AT@T, Faces94, UMIST, and FERET, databases. The accuracies for all databases when using 2D-WT are ranging between 96.18 % and 99.6 % except FERET and its accuracy is 92 %.     

Parallel probabilistic relaxation labelling based on Markov random fields for spectral-spatial hyperspectral image classification

The large volume of data and computational complexity of algorithms limit the application of hyperspectral image classification to real-time operations. This work addresses the use of different parallel processing techniques to speed up the Markov random field (MRF)-based method to perform spectral-spatial classification of hyperspectral imagery. The Metropolis relaxation labelling approach is modified to take advantage of multi-core central processing units (CPUs) and to adapt it to massively parallel processing systems like graphics processing units (GPUs). The experiments on different hyperspectral data sets revealed that the implementation approach has a huge impact on the execution time of the algorithm. The results demonstrated that the modified MRF algorithm produced classification accuracy similar to conventional methods with greatly improved computational performance. With modern multi-core CPUs, good computational speed-up can be achieved even without additional hardware support. The CPU-GPU hybrid framework rendered the otherwise computationally expensive approach suitable for time-constrained applications.     

3D卷积自编码器高光谱图像分类模型

目的 高光谱图像分类是遥感领域的基础问题，高光谱图像同时包含丰富的光谱信息和空间信息，传统模型难以充分利用两种信息之间的关联性，而以卷积神经网络为主的有监督深度学习模型需要大量标注数据，但标注数据难度大且成本高。针对现有模型的不足，本文提出了一种无监督范式下的高光谱图像空谱融合方法，建立了3D卷积自编码器（3D convolutional auto-encoder，3D-CAE）高光谱图像分类模型。方法 3D卷积自编码器由编码器、解码器和分类器构成。将高光谱数据预处理后，输入到编码器中进行无监督特征提取，得到一组特征图。编码器的网络结构为3个卷积块构成的3D卷积神经网络，卷积块中加入批归一化技术防止过拟合。解码器为逆向的编码器，将提取到的特征图重构为原始数据，用均方误差函数作为损失函数判断重构误差并使用Adam算法进行参数优化。分类器由3层全连接层组成，用于判别编码器提取到的特征。以3D-CNN （three dimensional convolutional neural network）为自编码器的主干网络可以充分利用高光谱图像的空间信息和光谱信息，做到空谱融合。以端到端的方式对模型进行训练可以省去复杂的特征工程和数据预处理，模型的鲁棒性和稳定性更强。结果 在Indian Pines、Salinas、Pavia University和Botswana等4个数据集上与7种传统单特征方法及深度学习方法进行了比较，本文方法均取得最优结果，总体分类精度分别为0.948 7、0.986 6、0.986 2和0.964 9。对比实验结果表明了空谱融合和无监督学习对于高光谱遥感图像分类的有效性。结论 本文模型充分利用了高光谱图像的光谱特征和空间特征，可以做到无监督特征提取，无需大量标注数据的同时分类精度高，是一种有效的高光谱图像分类方法。     

Hyperspectral image classification based on morphological profiles and decision fusion

Extended morphological profile (EMP) is an important mathematical tool for extracting structural information from the hyperspectral images. However, the accuracy of the EMP-based classification is greatly influenced by the choice of structuring element (SE). In this article, two supervised classification frameworks multiclassifier system with morphological profiles (MCSMP) and MCSMP2 are proposed that exploit rich spectral and structural information of hyperspectral images using EMPs and multiclassifier system for better classification than conventional methods. The EMPs with SEs of multiple shapes are used instead of one particular shape to better detect the response from the structures in the image. The EMPs created from SEs of different shapes are independently classified followed by decision fusion to generate final classification map. The classification results are compared with the conventional pixelwise and other EMP-based methods. The experimental results from three different types of hyperspectral data sets demonstrate that the proposed methods have significantly improved the spectral approach and outperformed the other studied methods in terms of classification accuracy. The new methods are more robust to the noise and produce good classification accuracy with very limited training samples. Various decision fusion techniques are evaluated, which performed differently in tested scenarios. Two different classifiers, Support Vector Machine (SVM) and random forest, are used in the experiments. It is shown that the proposed methods perform better with random forest classifier.     

Optimized discriminative transformations for speech features based on minimum classification error

Feature extraction is an important component of pattern classification and speech recognition. Extracted features should discriminate classes from each other while being robust to environmental conditions such as noise. For this purpose, several feature transformations are proposed which can be divided into two main categories: data-dependent transformation and classifier-dependent transformation. The drawback of data-dependent transformation is that its optimization criteria are different from the measure of classification error which can potentially degrade the classifier’s performance. In this paper, we propose a framework to optimize data-dependent feature transformations such as PCA (Principal Component Analysis), LDA (Linear Discriminant Analysis) and HLDA (Heteroscedastic LDA) using minimum classification error (MCE) as the main objective. The classifier itself is based on Hidden Markov Model (HMM). In our proposed HMM minimum classification error technique, the transformation matrices are modified to minimize the classification error for the mapped features, and the dimension of the feature vector is not changed. To evaluate the proposed methods, we conducted several experiments on the TIMIT phone recognition and the Aurora2 isolated word recognition tasks. The experimental results show that the proposed methods improve performance of PCA, LDA and HLDA transformation for mapping Mel-frequency cepstral coefficients (MFCC).     

Hyperspectral image compression based on online learning spectral features dictionary

Multimedia Tools and Applications - This paper proposes a novel method of lossy hyperspectral image compression using online learning dictionary. Spectral dictionary that learned in sparse coding...     

自适应迁移鲁棒特征的个性化联邦医学图像分类

目的 针对联邦学习中多中心医学数据的异质性特征导致全局模型性能不佳的问题,提出一种基于特征迁移的自适应个性化联邦学习算法（adaptive personalized federated learning via feature transfer, APFFT）。方法 首先,为降低全局模型中异质性特征信息影响,提出鲁棒特征选择网络（robust feature selection network, RFS-Net）构建个性化本地模型。RFS-Net通过学习两个迁移权重分别确定全局模型向本地模型迁移时的有效特征以及特征迁移的目的地,并构建基于迁移权重的迁移损失函数以加强本地模型对全局模型中有效特征的注意力,从而构建个性化本地模型。然后,为过滤各本地模型中异质性特征信息,利用自适应聚合网络（adaptive aggregation network, AANet）聚合全局模型。AA-Net基于全局模型交叉熵变化更新迁移权重并构建聚合损失,使各本地模型向全局模型迁移鲁棒特征,提高全局模型的特征表达能力。结果 在3种医学图像分类任务上与4种现有方法进行比较实验,在肺结核肺腺癌分类任务中,各中心曲线...     

基于频谱空间域特征注意的音乐流派分类算法

为了提升深度卷积神经网络对音乐频谱流派特征的提取效果,提出一种基于频谱空间域特征注意的音乐流派分类算法模型DCNN-SSA。DCNN-SSA模型通过对不同音乐梅尔谱图的流派特征在空间域上进行有效标注,并且改变网络结构,从而在提升特征提取效果的同时确保模型的有效性,进而提升音乐流派分类的准确率。首先,将原始音频信号进行梅尔滤波,以模拟人耳的滤波操作对音乐的音强及节奏变化进行有效过滤,所生成的梅尔谱图进行切割后输入网络;然后,通过深化网络层数、改变卷积结构及增加空间注意力机制对模型在流派特征提取上进行增强;最后,通过在数据集上进行多批次的训练与验证来有效提取并学习音乐流派特征,从而得到可以对音乐流派进行有效分类的模型。在GTZAN数据集上的实验结果表明,基于空间注意的音乐流派分类算法与其他深度学习模型相比,在音乐流派分类准确率和模型收敛效果上有所提高,准确率提升了5.36个百分点～10.44个百分点。     

Semi-supervised hyperspectral image classification algorithm based on graph embedding and discriminative spatial information

Image classification is one of the important techniques in computer vision. Due to the limited access of labeled samples in hyperspectral images, semi-supervised learning (SSL) methods have been widely applied in hyperspectral image classification. Graph based semi-supervised learning provides an effective solution to model data in classification problems, of which graph construction is the critical step. In this paper we employ the graphs constructed with a typical manifold learning method-locally linear embedding (LLE), based on which semi-supervised classification is then conducted. To exploit the valuable spatial information contained in hyperspectral images, discriminative spatial information (DSI) is then extracted. The proposed classification method is evaluated using three real hyperspectral data sets, revealing state-of-art performance when compared with different classification methods.     

Framework for robust blind image watermarking based on classification of attacks

With the widespread internet usage, digital contents are easily distributed throughout the world. To eliminate concerns of producers and owners of digital contents, watermarking techniques are extensively being used. Robustness against intentional and unintentional attacks is a major quality of watermarking systems. Since different attacks tend to target different parts of the frequency spectrum, in this paper we propose a framework for blind watermarking which determines the type of attack that the image has gone through before extracting the watermark. Within this framework, we propose an attack classification method to identify the region of the frequency spectrum that is less damaged. The watermark which is redundantly spread throughout the spectrum can be extracted from the less damaged regions. Experimental results show functionality of the framework by producing better results in comparison with well-known blind watermarking techniques.