基于焦点损失的半监督高光谱图像分类

针对高光谱图像（HSI）训练数据获取困难的问题,采用了一种新的HSI半监督分类框架,该框架利用有限的标记数据和丰富的未标记数据来训练深度神经网络。同时,由于高光谱样本分布是不平衡的,导致不同样本分类难度存在巨大差异,采用原始交叉熵损失函数无法刻画这种分布特征,因而分类效果不理想。为了解决这个问题,在半监督分类框架中提出一种基于焦点损失的多分类目标函数。最后,考虑到HSI的空间信息对分类的影响,结合马尔可夫随机场（MRF）,利用样本空间特征进一步改善分类效果。在两个常用的HSI数据集上,将所提方法与多种典型算法进行了实验对比分析,实验结果表明所提方法能够产生优于其他对比方法的分类效果。     

Semi-supervised dimension reduction based on hypergraph embedding for hyperspectral images

Dimension reduction (DR) is an efficient and effective preprocessing step of hyperspectral images (HSIs) classification. Graph embedding is a frequently used model for DR, which preserves some geometric or statistical properties of original data set. The embedding using simple graph only considers the relationship between two data points, while in real-world application, the complex relationship between several data points is more important. To overcome this problem, we present a linear semi-supervised DR method based on hypergraph embedding (SHGE) which is an improvement of semi-supervised graph learning (SEGL). The proposed SHGE method aims to find a projection matrix through building a semi-supervised hypergraph which can preserve the complex relationship of the data and the class discrimination for DR. Experimental results demonstrate that our method achieves better performance than some existing DR methods for HSIs classification and is time saving compared with the existed method SEGL which used simple graph.     

A Pre-processing framework for spectral classification of hyperspectral images

Multimedia Tools and Applications - Classification of Hyperspectral images is mostly based on the spectral-spatial features in existing classification techniques. The captured Hyperspectral images...     

A new residual fusion classification method for hyperspectral images

In this paper, we propose a novel residual fusion classification method for hyperspectral image using spatial–spectral information, abbreviated as RFC-SS. The RFC-SS method first uses the Gabor texture features and the non-parametric weighted spectral features to describe the hyperspectral image from both aspects of spatial and spectral information. Then it applies the residual fusion method to save the useful information from different classification methods, which can greatly improve the classification performance. Finally, the test sample is assigned to the class that has the minimal fused residuals. The RFC-SS classification method is tested on two classical hyperspectral images (i.e. Indian Pines, Pavia University). The theoretical analysis and experimental results demonstrate that the RFC-SS classification method can achieve a better performance in terms of overall accuracy, average accuracy, and the Kappa coefficient when compared to the other classification methods.     

Cloud removal for hyperspectral remotely sensed images based on hyperspectral information fusion

ABSTRACT

Hyperspectral remote sensing plays an important role in a wide variety of fields. However, its specific application for land surface analysis has been constrained due to the different shapes of thick, opaque cloud cover. The reconstruction of missing information obscured by clouds in remote-sensing images is an area of active research. However, most of the available cloud-removal methods are not suitable for hyperspectral images, because they lose the spectral information which is very important for hyperspectral analysis. In this article, we developed a new spectral resolution enhancement method for cloud removal (SREM-CR) from hyperspectral images, with the help of an auxiliary cloud-free multispectral image acquired at different times. In the fixed hyperspectral image, spectra of the cloud cover pixels are reconstructed depending on the relationship between the original hyperspectral and multispectral images. The final resulting image has the same spectral resolution as the original hyperspectral image but without clouds. This approach was tested on two experiments, in which the results were compared by visual interpretation and statistical indices. Our method demonstrated good performance.     

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.     

Caffe CNN-based classification of hyperspectral images on GPU

Deep learning techniques based on Convolutional Neural Networks (CNNs) are extensively used for the classification of hyperspectral images. These techniques present high computational cost. In this paper, a GPU (Graphics Processing Unit) implementation of a spatial-spectral supervised classification scheme based on CNNs and applied to remote sensing datasets is presented. In particular, two deep learning libraries, Caffe and CuDNN, are used and compared. In order to achieve an efficient GPU projection, different techniques and optimizations have been applied. The implemented scheme comprises Principal Component Analysis (PCA) to extract the main features, a patch extraction around each pixel to take the spatial information into account, one convolutional layer for processing the spectral information, and fully connected layers to perform the classification. To improve the initial GPU implementation accuracy, a second convolutional layer has been added. High speedups are obtained together with competitive classification accuracies.

     

基于监督局部线性嵌入特征提取的高光谱图像分类

高光谱图像的数据维数高、数据量大、数据间高度冗余等特点给图像分类带来困难,为进行有效降维、提高分类精度,提出了一种监督局部线性嵌入(SLLE)非线性流形学习特征提取方法。SLLE算法根据数据先验类标签信息所给出的新距离寻找数据点的k最近邻(NN),新距离使得类内距离小于类间距离,这使得SLLE算法更有利于分类。高光谱图像数据和UCI数据的分类结果表明了该方法的有效性。     

语义融合眼底图像动静脉分类方法

目的 眼底图像中的动静脉分类是许多系统性疾病风险评估的基础步骤。基于传统机器学习的方法操作复杂,且往往依赖于血管提取的结果,不能实现端到端的动静脉分类,而深度语义分割技术的发展使得端到端的动静脉分类成为可能。本文结合深度学习强大的特征提取能力,以提升动静脉分类精度为目的,提出了一种基于语义融合的动静脉分割模型SFU-Net（semantic fusion based U-Net）。方法 针对动静脉分类任务的特殊性,本文采用多标签学习的策略来处理该问题,以降低优化难度。针对动静脉特征的高度相似性,本文以DenseNet-121作为SFU-Net的特征提取器,并提出了语义融合模块以增强特征的判别能力。语义融合模块包含特征融合和通道注意力机制两个操作：1）融合不同尺度的语义特征从而得到更具有判别能力的特征;2）自动筛选出对目标任务更加重要的特征,从而提升性能。针对眼底图像中血管与背景像素之间分布不均衡的问题,本文以focal loss作为目标函数,在解决类别不均衡问题的同时重点优化困难样本。结果 实验结果表明,本文方法的动静脉分类的性能优于现有绝大多数方法。本文方法在DRIVE（digital retinal images for vessel extraction）数据集上的灵敏性（sensitivity）与目前最优方法相比仅有0.61%的差距,特异性（specificity）、准确率（accuracy）和平衡准确率（balanced-accuracy）与目前最优方法相比分别提高了4.25%,2.68%和1.82%;在WIDE数据集上的准确率与目前最优方法相比提升了6.18%。结论 语义融合模块能够有效利用多尺度特征并自动做出特征选择,从而提升性能。本文提出的SFU-Net在动静脉分类任务中表现优异,性能超越了现有绝大多数方法。     

A combination of k-means clustering and entropy filtering for band selection and classification in hyperspectral images

Hyperspectral images usually have large volumes of data comprising hundreds of spectral bands. Removal of redundant bands can both reduce computational time and improve classification performance. This work proposes and analyses a band-selection method based on the k-means clustering strategy combined with a classification approach using entropy filtering. Experimental results in different terrain images show that our method can significantly reduce the number of bands while maintaining an accurate classification.     

基于离散度量的半监督分类算法

半监督分类算法试图根据已知样本对特定的未知样本建立一套进行识别的方法和准则。渐进直推式分类学习算法是一种基于SVM的半监督分类学习方法,在基于渐进直推式分类学习算法的基础上,利用Fisher准则中的样本离散度作为度量标准,采用Fisher准则函数作为评价函数,提出了一种基于离散度量和SVM相结合的半监督分类算法,在时间复杂度和样本测试精度上较PTSVM算法都取得了良好的学习效果。     

Efficient classification of the hyperspectral images using deep learning

Multimedia Tools and Applications - Classification techniques applicable to the hyperspectral images do not extract deep features from the hyperspectral image efficiently. In this work, a deep...     

基于权值多样性的半监督分类算法

在实际生活中,可以很容易地获得大量系统数据样本,却只能获得很小一部分的准确标签.为了获得更好的分类学习模型,引入半监督学习的处理方式,对基于未标注数据强化集成多样性(UDEED)算法进行改进,提出了UDEED+——一种基于权值多样性的半监督分类算法.UDEED+主要的思路是在基学习器对未标注数据的预测分歧的基础上提出权...     

基于相对变换的半监督分类算法

为了增强基于图的局部和全部一致性(LGC)半监督算法的处理稀疏和噪声数据的能力,提出了一种基于相对变换的LGC算法。该算法通过相对变换将原始数据空间转换到相对空间,在相对空间中噪声和孤立点远离正常点,稀疏的数据变得相对密集,从而可以提高算法的性能。仿真实验结果表明,基于相对变换的LGC算法有更强的处理稀疏和噪声数据的能力。     

Compression and classification methods for hyperspectral images

In this article we present new lossless compression methods by combining existing methods and compare them using AVIRIS images. These methods include the Self-Organizing Map (SOM), Principal Component Analysis (PCA), and the three-dimensional Wavelet Transform combined with traditional lossless encoding methods. The two-dimensional JPEG2000 and SPIHT compression methods were applied to the eigenimages produced by the PCA. The bit allocation for the compression of eigenimages was based on the amount of information in each eigenimage. In bit rate calculation we used the exponential entropy formula, which gave better results than the original linear version. The information loss from the compression was measured by the Signal-to-Noise Ratio (SNR) and Peak-Signal-to-Noise Ratio (PSNR). To get more illustrative and practical error measures, classification of spectra was performed using unsupervised K-means clustering combined with spectral matching. Spectral matching methods include Euclidean distance, Spectral Similarity Value (SSV), and Spectral Angle Mapper (SAM). We used two test images, which both were AVIRIS images with 224 bands and 512 lines in 614 columns. The PCA in the spectral dimension combined with JPEG2000 or SPIHT in the spatial dimension was the best method in terms of the image quality and compression speed.     

An adaptive feature fusion framework for multi-class classification based on SVM

An adaptive feature fusion framework is proposed for multi-class classification based on SVM. In a similar manner of one-versus-all (OVA), one of the multi-class SVM schemes, the proposed approach decomposes a multi-class classification into several binary classifications. The main difference lies in that each classifier is created with the most suitable feature vectors to discriminate one class from all the other classes. The feature vectors of the unknown samples are selected by each classifier adaptively such that recognition is fulfilled accordingly. In addition, novel evaluation criterions are defined to deal with the frequent small-number sample problems. A writer recognition experiment is carried out to accomplish this framework with three kinds of feature vectors: texture, structure and morphological features. Finally, the performance of the proposed approach is illustrated as compared with the OVA by applying the same feature vectors for all classes.     

ICA-based fusion for colour display of hyperspectral images

Hyperspectral images contain data from a large number of contiguous bands and, therefore, cannot be displayed directly using a colour display system. In this paper, an independent component analysis-based (ICA-based) approach for the problem of fusing hyperspectral images to three-band images for colour display purposes is proposed. Correlation coefficient and mutual information (ICA-CCMI) are used as criteria for selecting three suitable independent components for colour representation. In addition, statistical evaluation metrics for the colour display results of hyperspectral images are provided and discussed in light of different visualization goals. A new quality metric motivated by the quality index is developed to evaluate the structural information of the colour display images. The performance of our approach is validated by applying it to three hyperspectral image datasets. The experimental results demonstrate promising performance for the ICA-CCMI algorithm, compared with existing principal component analysis-based (PCA-based) methods for visualization of hyperspectral images.     

基于辅助空间与极大熵的半监督聚类方法

将极大熵原理引入半监督聚类方法中,提出基于辅助空间与极大熵的半监督聚类算法AMESC,针对该算法中的代价函数进行迭代优化,实现聚类。AMESC的优势在于它依据模拟退火过程,使算法避开局部极小而得到全局极小,提高算法性能。通过实验证实了AMESC的有效性和优越性。