Rotation-invariant texture classification using a two-stage waveletpacket feature approach

A novel two-stage wavelet packet feature approach for classification of rotated textured images is discussed. In the first stage, a set of sorted and dominant wavelet packet features is extracted from a texture image and a Mahalanobis distance classifier is employed to output N best classes. In the second stage, another set of wavelet packet features is extracted from the polarised form of the sample texture image and the most dominant wavelet packet features are selected and passed to the radial basis function (RBF) classifier with the N best classes to output the final matched class. Experimental results, based on a large sample data set of twenty distinct natural textures selected from the Brodatz album with different orientations, show that the proposed method outperforms the similar wavelet methods and the other rotation invariant texture classification schemes, and an overall accuracy rate of 91.4% was achieved     

基于多特征融合的层次支持向量机遥感图像云检测

云检测是遥感图像处理和应用的前提,针对遥感图像云检测的准确率容易受到薄云及似云地物影响的挑战,提出一种结合遥感影像灰度、纹理和频率特征的层次支持向量机云检测算法.该方法首先采用简单线性迭代聚类算法将遥感图像分割为像素块,再采用一种层次支持向量机分类器对遥感图像以像素块为单位进行云检测.层次支持向量机的第一层将像素块初步...     

SWS-DAN: Subtler WS-DAN for fine-grained image classification

Currently, weakly supervised data augmentation network (WS-DAN) has been proved to be one of the state-of-the-art methods for fine-grained image classification due to its effectiveness on attention-guided data augmentation and bilinear attention pooling. Taking WS-DAN as the backbone, in this paper, we further propose a subtler WS-DAN recognition network, namely, SWS-DAN. Specifically, we first construct a novel “salience-guided data augmentation” scheme composed of cutblock, part-aware cropping, and SCutMix operations, which can more effectively expand the number of training dataset and improve the weakness addressed in the data augmentation procedure of WS-DAN. Meanwhile, the novel data-augmentation manner reduces background noise and mines more discriminative regions simultaneously, thereby avoiding the overfitting. In caring about the key issue in fine-grained image classification task is how to distinguish the extremely similar subclasses (e.g., Artic Tern, Elegant Tern, and Forsters Tern), we then design a “Top-k” loss function that mainly focuses on the similar classes so as to find their extraordinary subtle differences. Extensive experiments carried out on common fine-grained image datasets demonstrate that SWS-DAN can surpass WS-DAN with a significant margin in the classification performance.     

Two-stage multiscan detection of target using the “l/n - d” criterion with identification of pips by the maximum of decision statistic

Algorithms of two-stage multiscan detection of target based on the incoherent accumulation using the “l/n - d” criterion have been obtained. The primary detection is performed at the first stage with sufficiently high probability of false alarm. At the second stage the tracking task is solved by using the “strongest neighbor” criterion and the incoherent interscan accumulation of decision statistics is performed. Expressions for probabilities of the true detection of target and false alarm of the algorithm using the “3/5 - 2” criterion were obtained. This algorithm was analyzed by using the statistical simulation for a case of multiscan target detection from data of the radar measuring the range and radial velocity of target.     

Multiple kernel-based multi-instance learning algorithm for image classification

In this paper, a novel multi-instance learning (MIL) algorithm based on multiple-kernels (MK) framework has been proposed for image classification. This newly developed algorithm defines each image as a bag, and the low-level visual features extracted from its segmented regions as instances. This algorithm is started from constructing a “word-space” from instances based on a collection of “visual-words” generated by affinity propagation (AP) clustering method. After calculating the distance between a “visual-word” and the bag (image), a nonlinear mapping mechanism is introduced for registering each bag as a coordinate point in the “word-space”. In this case, the MIL problem is transformed into a standard supervised learning problem, which allows multiple-kernels support vector machine (MKSVM) classifiers to be trained for the image categorization. Compared with many popular MIL algorithms, the proposed method, named as MKSVM-MIL, shows its satisfactorily experimental results on the COREL dataset, which highlights the robustness and effectiveness for image classification applications.     

Two-stage salient region detection by exploiting multiple priors

Saliency detection has become a valuable tool for many image processing tasks, like image retargeting, object recognition, and adaptive compression. With the rapid development of the saliency detection methods, people have approved the hypothesis that “the appearance contrast between the salient object and the background is high”, and build their saliency methods on some priors that explain this hypothesis. However, these methods are not satisfactory enough. We propose a two-stage salient region detection method. The input image is first segmented into superpixels. In the first stage, two measures which measure the isolation and distribution of each superpixel are proposed, we consider that both of these two measures are important for finding the salient regions, thus the image-feature-based saliency map is obtained by combining the two measures. Then, in the second stage, we incorporate into the image-feature-based saliency map a location prior map to emphasize the foci of attention. In this algorithm, six priors that explain what is the salient region are exploited. The proposed method is compared with the state-of-the-art saliency detection methods using one of the largest publicly available standard databases, the experimental result indicates that the proposed method has better performance. We also demonstrate how the saliency map of the proposed method can be used to create high quality of initial segmentation masks for subsequent image processing, like Grabcut based salient object segmentation.     

Anomaly Detection Techniques in the Gaia Space Mission Data

In this paper we deal with classification of anomalous data detected by the data reduction system of the Gaia space mission, in operation since 2013. Given the size and complexity of intermediate data and plots for diagnostics, beyond practical possibility of full human evaluation, the need for automated signal processing tools is becoming more and more relevant. Our classification task consists in discriminating among “normal” data and data affected by anomalies, which at present are grouped into four different classes. We investigate the use of some clever pre-processing approaches that allow the application of a tailored technique based on the Hough transform, and of some machine learning tools, evidencing that the task can be exactly solved in the former case. In the latter case, random forests and support vector machine provide less than satisfactory performance, while convolutional neural networks achieve very good classification accuracy, up to 91.22%. Statistics show satisfactory results also in terms of precision and recall of each class.

     

Interframe Predictive Coding of Images Using Hybrid Vector Quantization

The temporal and spatial redundancies of image sequences can be reduced by prediction and vector quantization, respectively. The robustness of the coder is increased through the use of a hybrid two-stage vector quantizer. The first stage is of the Linde-Buzo-Gray type. The second stage encodes the residual error of the first stage using a lattice vector quantizer.     

A pixel-based approach for classification of cardiac single photon emission computed tomography images

In this work, a computer-based algorithm is proposed for the initial interpretation of human cardiac images. Reconstructed single photon emission computed tomography images are used to differentiate between subjects with normal value and abnormal value of ejection fraction. The method analyses pixel intensities that correspond to blood flow in the left ventricular region. The algorithm proceeds through three main stages: the initial stage does a pre-processing task to reduce noise as well as blur in the image. The second stage extracts features from the images. Classification is done in the final stage. The pre-processing stage consists of a de-noising part and a de-blurring part. Novel features are used for classification. Features are extracted as three different sets based on: the pixel intensity distribution in different regions, spatial relationship of pixels and multi-scale image information. Two supervised algorithms are proposed for classification: one algorithm is based on a threshold value computed from the features extracted from the training images and the other algorithm is based on sequential minimal optimization-based support vector machine approach. Experimental studies were performed on real cardiac SPECT images obtained from hospital. The result of classification has been verified by an expert nuclear medicine physician and by the ejection fraction value obtained from quantitative gated SPECT, the most widely used software package for quantifying gated SPECT images.     

基于稀疏约束的SAR目标特征提取方法研究

特征提取是合成孔径雷达目标识别关键技术与核心任务。为了更好地提取目标特征,稀疏约束将被添加在非负矩阵分解法中,并应用于图像目标特征提取,通过利用稀疏约束的非负矩阵分解方法对sAR目标图像进行分解,构建具有稀疏性的目标特征矢量,提高了特征矢量的类内相似性与类间差异性。利用基于支持向量机的分类方法对MSTAR数据进行目标识别试验,试验结果表明,添加稀疏约束的NMF方法与PCA、ICA以及一般NMF特征提取方法相比,能够显著提高目标识别的稳定性和准确率。     

Application of neural networks for sea ice classification inpolarimetric SAR images

Several automatic methods have been developed to classify sea ice types from fully polarimetric synthetic aperture radar (SAR) images, and these techniques are generally grouped into supervised and unsupervised approaches. In previous work, supervised methods have been shown to yield higher accuracy than unsupervised techniques, but suffer from the need for human interaction to determine classes and training regions. In contrast, unsupervised methods determine classes automatically, but generally show limited ability to accurately divide terrain into natural classes. In this paper, a new classification technique is applied to determine sea ice types in polarimetric and multifrequency SAR images, utilizing an unsupervised neural network to provide automatic classification, and employing an iterative algorithm to improve the performance. The learning vector quantization (LVQ) is first applied to the unsupervised classification of SAR images, and the results are compared with those of a conventional technique, the migrating means method. Results show that LVQ outperforms the migrating means method, but performance is still poor. An iterative algorithm is then applied where the SAR image is reclassified using the maximum likelihood (ML) classifier. It is shown that this algorithm converges, and significantly improves classification accuracy. The new algorithm successfully identifies first-year and multiyear sea ice regions in the images at three frequencies. The results show that L- and P-band images have similar characteristics, while the C-band image is substantially different. Classification based on single features is also carried out using LVQ and the iterative ML method. It is found that the fully polarimetric classification provides a higher accuracy than those based on a single feature. The significance of multilook classification is demonstrated by comparing the results obtained using four-look and single-look classifications     

Combining pores and ridges with minutiae for improved fingerprint verification

This paper presents a fast fingerprint verification algorithm using level-2 minutiae and level-3 pore and ridge features. The proposed algorithm uses a two-stage process to register fingerprint images. In the first stage, Taylor series based image transformation is used to perform coarse registration, while in the second stage, thin plate spline transformation is used for fine registration. A fast feature extraction algorithm is proposed using the Mumford–Shah functional curve evolution to efficiently segment contours and extract the intricate level-3 pore and ridge features. Further, Delaunay triangulation based fusion algorithm is proposed to combine level-2 and level-3 information that provides structural stability and robustness to small changes caused due to extraneous noise or non-linear deformation during image capture. We define eight quantitative measures using level-2 and level-3 topological characteristics to form a feature supervector. A 2ν-support vector machine performs the final classification of genuine or impostor cases using the feature supervectors. Experimental results and statistical evaluation show that the feature supervector yields discriminatory information and higher accuracy compared to existing recognition and fusion algorithms.     

Classified vector quantisation with variable block-size DCT models

The paper describes the classified vector quantisation (CVQ) of an image, based on quadtrees and a classification technique in the discrete cosine transform (DCT) domain. In this scheme, a quadtree is used to segment low-detail regions into variable sized blocks and high-detail regions into uniform 4×4 blocks of various edge and mixed classes. High-detail blocks are classified by an edge-oriented classifier which employs a pattern-matching technique with edge models defined in the normalised DCT domain. The proposed classifier is simple to implement, and efficiently classifies edges to good visual accuracy. The low-detail regions are encoded at very low bit rates with little perceptual degradation, while the encoding of the high-detail regions is performed to achieve a good perceptual quality in the decoded image. Decoded images of high visual quality are obtained for encoding rates between 0.3 and 0.7 bpp     

Results and implications of a study of fifteen years of satellite image classification experiments

A study has been carried out of 15 years of published peer-reviewed experiments on satellite image classification. The aim of the study was to assess the degree of progress being made in thematic mapping through developments in classification algorithms and also in systems approaches such as postclassification analysis, multiclassifier integration, and data fusion. The results of over 500 reported classification experiments were quantitatively analyzed. This involved examination of relationships between classification accuracy and date of publication, as well as between accuracy and various experimental parameters such as number of classes, size of feature vector, resolution of satellite data, and test area. Comparisons were also made between different types of methodology such as neural network and nonneural approaches. Overall, the results show that there has been no demonstrable improvement in classification performance over the 15-year period. The mean value of the Kappa coefficient across all experiments was found to be 0.6557 with a standard deviation of 0.1980. Expected relationships between classification accuracy and resolution and between accuracy and number of classes were also not observed in the data. Some of the implications of these findings for the future research agenda are considered.     

基于图像配准的混合遗传FCM算法研究

针对目前图像变化检测的相关研究,提出一种新的算法:基于SAR图像配准的混合遗传FCM算法.算法主要分为4个步骤.第一步,利用Harris算法和SIFT算法对两幅图像进行匹配,证明它们是同源不同时相的图像.第二步,利用两种不同变化检测方法提取初步差异图像.第三步,利用PCA方法对差异图像进行降维处理.第四步,利用混合遗传FCM算法对特征矢量空间进行分类,并将分类结果与参考差异图像进行比较,获得变换信息.采用渥太华地区的部分图像作为检测算法的性能的数据库.获得的结果与FCM算法相比较,结果表明,提出的算法具有最高的全局正确率98.10％,算法效果更佳.     

Semantic clustering for region-based image retrieval

With the proliferation of applications that demand content-based image retrieval, two merits are becoming more desirable. The first is the reduced search space, and the second is the reduced “semantic gap.” This paper proposes a semantic clustering scheme to achieve these two goals. By performing clustering before image retrieval, the search space can be significantly reduced. The proposed method is different from existing image clustering methods as follows: (1) it is region based, meaning that image sub-regions, instead of the whole image, are grouped into. The semantic similarities among image regions are collected over the user query and feedback history; (2) the clustering scheme is dynamic in the sense that it can evolve to include more new semantic categories. Ideally, one cluster approximates one semantic concept or a small set of closely related semantic concepts, based on which the “semantic gap” in the retrieval is reduced.     

Principal component analysis-enhanced cosine radial basis function neural network for robust epilepsy and seizure detection.

A novel principal component analysis (PCA)-enhanced cosine radial basis function neural network classifier is presented. The two-stage classifier is integrated with the mixed-band wavelet-chaos methodology, developed earlier by the authors, for accurate and robust classification of electroencephalogram (EEGs) into healthy, ictal, and interictal EEGs. A nine-parameter mixed-band feature space discovered in previous research for effective EEG representation is used as input to the two-stage classifier. In the first stage, PCA is employed for feature enhancement. The rearrangement of the input space along the principal components of the data improves the classification accuracy of the cosine radial basis function neural network (RBFNN) employed in the second stage significantly. The classification accuracy and robustness of the classifier are validated by extensive parametric and sensitivity analysis. The new wavelet-chaos-neural network methodology yields high EEG classification accuracy (96.6%) and is quite robust to changes in training data with a low standard deviation of 1.4%. For epilepsy diagnosis, when only normal and interictal EEGs are considered, the classification accuracy of the proposed model is 99.3%. This statistic is especially remarkable because even the most highly trained neurologists do not appear to be able to detect interictal EEGs more than 80% of the times.     

Cross-layer progressive attention bilinear fusion method for fine-grained visual classification

Fine-grained visual classification (FGVC) is a critical task in the field of computer vision. However, FGVC is full of challenges due to the large intra-class variation and small inter-class variation of the classes to be classified on an image. The key in dealing with the problem is to capture subtle visual differences from the image and effectively represent the discriminative features. Existing methods are often limited by insufficient localization accuracy and insufficient feature representation capabilities. In this paper, we propose a cross-layer progressive attention bilinear fusion (CPABF in short) method, which can efficiently express the characteristics of discriminative regions. The CPABF method involves three components: 1) Cross-Layer Attention (CLA) locates and reinforces the discriminative region with low computational costs; 2) The Cross-Layer Bilinear Fusion Module (CBFM) effectively integrates the semantic information from the low-level to the high-level 3) Progressive Training optimizes the parameters in the network to the best state in a delicate way. The CPABF shows excellent performance on the four FGVC datasets and outperforms some state-of-the-art methods.     

Localized image watermarking in spatial domain resistant to geometric attacks

Watermark robustness to geometric attacks is still a challenging research field. In this paper, a novel robust image watermarking scheme is proposed for resisting such attacks. Watermark synchronization is first achieved by local invariant regions which can be generated using scale normalization and image feature points. The watermark is embedded into all the local regions repeatedly in spatial domain. During embedding, each circular region is first divided into homocentric cirque regions. Then the watermark bits are embedded by quantizing each cirque region into an “odd” or “even” region using odd–even quantization. In the decoder, an odd–even detector (OED) is designed to extract the watermark from the distorted image directly. Localized embedding achieves good invisibility and repeated insertion enhances watermark robustness. Simulation results show that the proposed scheme is robust to both geometric attacks and traditional signal processing attacks.     

Textural Infornation in SAR Images

A multiplicative model was used to relate the image variance for a given land-use category to the individual variances associated with image speckle and target texture. Speckle was treated as a random process governed by signal fading and was considered to be statistically independent of the textural variations associated with the spatial variations of the scattering properties of visually "uniform" distributed targets. Seasat SAR imagery of Oklahoma was used to evaluate the textural autocorrelation function of five land-use categories: water, forest, pasture, urban, and cultivated. It was found that the maximum classification accuracy achievable using first-order statistics was 72 percent and that this level of accuracy was obtainable only by significantly degrading the spatial resolution in order to increase the number of independent samples per pixel. In contrast, second-order statistics-specifically, image contrast and inverse moment-provided a classification accuracy of 88 percent, with only a modest degradation in spatial resolution. A second study using SIR-A imagery of five forested regions has shown that the use of textural information can improve the classification accuracy among the five forest types from 75 to 93 percent.