The field of meta-learning has seen a dramatic rise in interest in recent years. In existing meta-learning approaches, learning tasks for training meta-models are usually collected from public datasets, which brings the difficulty of obtaining a sufficient number of meta-learning tasks with a large amount of training data. In this paper, we propose a meta-learning approach based on randomly generated meta-learning tasks to obtain a parametric loss for classification learning based on big data. The loss is represented by a deep neural network, called meta-loss network (MLN). To train the MLN, we construct a large number of classification learning tasks through randomly generating training data, validation data, and corresponding ground-truth linear classifier. Our approach has two advantages. First, sufficient meta-learning tasks with large number of training data can be obtained easily. Second, the ground-truth classifier is given, so that the difference between the learned classifier and the ground-truth model can be measured to reflect the performance of MLN more precisely than validation accuracy. Based on this difference, we apply the evolutionary strategy algorithm to find out the optimal MLN. The resultant MLN not only leads to satisfactory learning effects on generated linear classifier learning tasks for testing, but also behaves very well on generated nonlinear classifier learning tasks and various public classification tasks. Our MLN stably surpass cross-entropy (CE) and mean square error (MSE) in testing accuracy and generalization ability. These results illustrate the possibility of achieving satisfactory meta-learning effects using generated learning tasks.

A "sign" on a lung CT image refers to a radiologic finding that suggests a pathological progression of some specific disease. Analysis of CT signs is helpful to understand the pathological origin of the lesion. In-depth study of lung nodules classification with different CT signs will help to distinguish benign and malignant nodules more clearly and accurately. To this end, we propose an Inception module-based ensemble classification method for pulmonary nodule diagnosis with different nodule signs. We first construct a Convolutional Neural Network (CNN) classifier adopting Inception modules and pre-train it on ImageNet. We then fine-tune this pre-trained classifier on 10 different lung nodule sign sample sets, and fuse these 10 classifiers with an artificial immune ensemble algorithm. The overall sensitivity, specificity, and accuracy of our proposed Artificial Immune Algorithm-based Inception Networks Fusion (AIA-INF) algorithm are 82.22%, 93.17%, and 88.67%, respectively, which are significantly higher than those of the alternative Bagging and Boosting methods. The experimental results show that our Inception-based ensemble classifier offers promising performance, and compared with other CADx systems, this scheme can offer a more detailed reference for diagnosis, and can be valuable for junior radiologist training.

For content-based image retrieval, the shape is one of the most important discriminatory elements. The form captures most of the perceptual information of the observed objects on images in many applications, while colour and texture can often be omitted without affecting the performance of the retrieval. Unfortunately, there may be significant changes in shape, such as deformation, scaling, changes in orientation noise, and partial concealment. Accurate shape description remains, therefore, a challenging technical issue. The study performs experimental analysis to identify the problem. The adoption of the MPEG-7 and KIMIA-99 standard has significant importance to simplify the image retrieval process. The Fourier Descriptors, Moment-Based Features, Hierarchical Centroids and Histogram of Oriented Gradients have been applied forextraction of images from datasets. The fusion of features has been done by Discriminant Correlation Analysis and Direct Concatenation of features it has been evident that by fusion of features we obtained approximately 90% accurate and better results.

Purpose. Computer-aided diagnosis (CAD) can aid in improving diagnostic level; however, the main problem currently faced by CAD is that it cannot obtain sufficient labeled samples. To solve this problem, in this study, we adopt a generative adversarial network (GAN) approach and design a semisupervised learning algorithm, named G2C-CAD. Methods. From the National Cancer Institute (NCI) Lung Image Database Consortium (LIDC) dataset, we extracted four types of pulmonary nodule sign images closely related to lung cancer: noncentral calcification, lobulation, spiculation, and nonsolid/ground-glass opacity (GGO) texture, obtaining a total of 3,196 samples. In addition, we randomly selected 2,000 non-lesion image blocks as negative samples. We split the data 90% for training and 10% for testing. We designed a DCGAN generative adversarial framework and trained it on the small sample set. We also trained our designed CNN-based fuzzy Co-forest on the labeled small sample set and obtained a preliminary classifier. Then, coupled with the simulated unlabeled samples generated by the trained DCGAN, we conducted iterative semisupervised learning, which continually improved the classification performance of the fuzzy Co-forest until the termination condition was reached. Finally, we tested the fuzzy Co-forest and compared its performance with that of a C4.5 random decision forest and the G2C-CAD system without the fuzzy scheme, using ROC and confusion matrix for evaluation. Results. Four different types of lung cancer-related signs were used in the classification experiment: noncentral calcification, lobulation, spiculation, and nonsolid/ground-glass opacity (GGO) texture, along with negative image samples. For these five classes, the G2C-CAD system obtained AUCs of 0.946, 0.912, 0.908, 0.887, and 0.939, respectively. The average accuracy of G2C-CAD exceeded that of the C4.5 random decision tree by 14%. G2C-CAD also obtained promising test results on the LISS signs dataset; its AUCs for GGO, lobulation, spiculation, pleural indentation, and negative image samples were 0.972, 0.964, 0.941, 0.967, and 0.953, respectively. Conclusion. The experimental results show that G2C-CAD is an appropriate method for addressing the problem of insufficient labeled samples in the medical image analysis field. Moreover, our system can be used to establish a training sample library for CAD classification diagnosis, which is important for future medical image analysis.

The automatic recognition of cavity imaging signs in lung computed tomography (CT) images is of great importance for early diagnosis and possible cure of lung tuberculosis and cancers. The performance of existing recognition methods which adopt classical technology needs to be improved accordingly. In this paper, we propose an automatic recognition method based on hybrid resampling and multi-feature fusion strategies. The hybrid resampling includes multi-receptive-field and multi-window settings: the former reduces the risk of missing small or large cavities, the latter reserves context information of multiply CT windows more compactly. For multi-feature fusion, we extract features of convolutional neural networks (CNN) and classical methods (histograms of oriented gradients (HOG) and local binary pattern (LBP)). Then we compress CNN-HOG features by principal components analysis (PCA) algorithm and combine them with LBP feature. Finally, we use the fused feature to train a support vector machine (SVM) model for improving classification performance. We evaluate our method on the cavity samples from LIDC-IDRI and LISS publicly available dataset of chest CT scans, which contains 167 cavities in 164 CT images. The experimental results show that fused feature has better discriminative capability than any single feature, and has the highest FS score (0.1472 vs 0.1136) in the group with sensitivity greater than 0.8. The proposed method is compared with the latest methods for the automatic recognition of cavity imaging sign and enables higher sensitivity than the second-best method (85% vs 70%). The experiment shows that the fusion of CNN feature and classical hand-crafted feature makes full use of the complementary information, and improves classification performance when number of samples in our application is limited.

Content-sensitive superpixel segmentation generates small superpixels in content-dense regions and large superpixels in content-sparse regions. It achieves higher segmentation accuracy than traditional superpixels. In this paper, we propose a content-sensitive superpixel segmentation algorithm based on Self-Organization-Map (SOM) neural network. First, we propose a novel metric to measure the content-sensitiveness of superpixels. Second, by using this metric, we develop a sampling algorithm to sample pixels from image according to their content-sensitiveness. Finally, a SOM neutral network is trained with the sampled pixels and used to segment the image into content-sensitive superpixels. The Berkeley Image Segmentation database and INRIA database are used to evaluate the proposed method. The experiment results show that the proposed approach outperforms state-of-the-art methods. (C) 2019 Published by Elsevier Inc.

Cloud-based storage services are multiplying and are being adapted mostly for data storage. At the same time, many potential problems pertaining to data storage and security are being addressed. This research provides the architecture for splitting user data and the solution in retrieving different data chunks, stored on different cloud storages. By doing this, not only the load on a single server is reduced but security and storage are efficiently used. Data would be stored and retrieved in slices, hence the chances of data forgery are diminishing. File processing will be faster as split parts would be fetched from different clouds, enabling parallel processing.

In this paper, an efficient and accurate multilevel change detection method is presented, achieved by the combination of an unsupervised object-based correlation analysis and a supervised post-classification comparison. Before the change detection procedure, fast multitemporal segmentation is applied to provide object-level information on two registered images. Then the proposed object-based correlation analysis method is used to extract the potential changed areas efficiently and stably, which improves the accuracy of overall performance. Notably, all the procedures are highly automatic except for the necessary selection of training examples within all supervised algorithms. The experimental results demonstrate the superior performance of our method compared with the four typical state-of-the-art change detection methods.

Various superpixel approaches have been published recently. These algorithms are assessed using different evaluation metrics and datasets resulting in discrepancy in algorithm comparison. This calls for a benchmark to compare the state-of-the-arts methods and evaluate their pros and cons. We analyze benchmark metrics, datasets and built a superpixel benchmark. We evaluated and integrated top 15 superpixel algorithms, whose code are publicly available, into one code library and, provide a quantitative comparison of these algorithms. We find that some superpixel algorithms perform consistently better than others. Clustering based superpixel algorithms are more efficient than graph-based ones. Furthermore, we also introduced a novel metric to evaluate superpixel regularity, which is a property that superpixels desired. The evaluation results demonstrate the performance and limitations of state-of-the-art algorithms. Our evaluation and observations give deep insight about different algorithms and will help researchers to identify the more feasible superpixel segmentation methods for their different problems.

Sea-land segmentation is a key step for some important applications of panchromatic remote sensing image processing. However, robust and effective sea-land segmentation for high-resolution panchromatic remote sensing images is still a challenging problem. This letter presents an accurate and robust approach by integrating the improved multiscale normalized cut (IMNcut) method and improved Chan-Vese model for sea-land segmentation. At first, the image is downsampled and segmented into multiple regions by the IMNcut method. Next, the homogeneous regions are merged to obtain a coarse segmentation result. Finally, gray intensity and local entropy features are integrated as discriminants of the improved Chan-Vese model, which is used to obtain the final segmentation result through a low- to high-resolution segmentation scheme. Experimental results performed on several real data sets demonstrate the effectiveness of the proposed model in terms of visual and objective evaluations.

>Dear editor,Middle wave infrared remote(MWIR)technology is the main approach in remote sensing applications,because thermal infrared can work all-time.The MWIR technology can provide valid object information in remote sensing technology than optical and synthetic aperture radar(SAR)images

Computer-aided detection (CAD) of lobulation can help radiologists to diagnose/detect lung diseases easily and accurately. Compared to CAD of nodule and other lung lesions, CAD of lobulation remained an unexplored problem due to very complex and varying nature of lobulation. Thus, many state-of-the-art methods could not detect successfully. Hence, we revisited classical methods with the capability of extracting undulated characteristics and designed a sliding window based framework for lobulation detection in this paper. Under the designed framework, we investigated three categories of lobulation classification algorithms: template matching, feature based classifier, and bending energy. The resultant detection algorithms were evaluated through experiments on LISS database. The experimental results show that the algorithm based on combination of global context feature and BOF encoding has best overall performance, resulting in F1 score of 0.1009. Furthermore, bending energy method is shown to be appropriate for reducing false positives. We performed bending energy method following the LIOP-LBP mixture feature, the average positive detection per image was reduced from 30 to 22, and F1 score increased to 0.0643 from 0.0599. To the best of our knowledge this is the first kind of work for direct lobulation detection and first application of bending energy to any kind of lobulation work.

Automatic sea-land segmentation is an essential and challenging field for the practical use of panchromatic satellite imagery. Owing to the temporal variations as well as the complex and inconsistent intensity contrast in both land and sea areas, it is difficult to generate an accurate segmentation result by using the conventional thresholding methods. Additionally, the freely available digital elevation model (DEM) also difficultly meets the requirements of high-resolution data for practical usage, because of the low precision and high memory storage costs for the processing systems. In this case, we proposed a fully automatic sea-land segmentation approach for practical use with a hierarchical coarse-to-fine procedure. We compared our method with other state-of-the-art methods with real images under complex backgrounds and conducted quantitative comparisons. The experimental results show that our method outperforms all other methods and proved being computationally efficient.

Automatic detection of moving targets is one of important research area in the remote sensing field. In this paper, we propose a method that accurately detects moving targets in aerial videos using hierarchical spatiotemporal saliency analysis. First, coarse motion regions are extracted by utilizing global temporal saliency analysis. Based on these local candidate regions, spatial saliency methods are used to obtain accurate description of targets. After fusing spatial and temporal saliency values, we can get refined results of the detection. Considering about the inter-frame consistency of motion, trajectory level analysis is added in the proposed method to eliminate false alarms. Experiments conducted on the VIVID dataset validate the effectiveness and efficiency of the proposed method.

Traditional region-of-interest (ROI) detection methods for remote sensing images are generally formulated at pixel level and are less efficient when applied on large high-resolution images. This letter presents an accurate and efficient approach via superpixel-to-pixel saliency analysis for ROI detection. At first, the image is downsampled and segmented into superpixels by simple linear iterative clustering. Next, structure tensor and background contrast are used to yield superpixel feature maps for texture and color. After fusing the feature maps, the overall superpixel saliency map is obtained and then used to achieve the final pixel-level saliency map by superpixel-to-pixel mapping. Through experimentations, we validate the effectiveness and computational efficiency of the proposed model in comparison with state-of-the-art techniques.

Ground-glass opacity (GGO) detection is paramount for the prognosis and diagnosis of lung diseases. We present a novel GGO detection method for 2D lung CT images in this paper, which focuses on detecting GGOs with high sensitivity and reducing false positives as much as possible. To this end, we propose a local-to-global multilevel thresholding algorithm for segmentation and a novel discriminative learning algorithm for identification to solve the problem of GGO detection. There are two components in our method. In the first component, we perform clustering on the local Ostu thresholds of CT levels for each patch of an image, the candidate regions of interests (ROIs) are segmented based on the clustering results by multilevel thresholding techniques. The second component is a Bayesian modeling process for identifying the GGOs from ROI candidates, the classifier is trained based on Bayesian risk minimization and margin maximization by our discriminative learning algorithm. The proposed GGO detection approach is evaluated on the LISS database with 45 GGOs. Finally, our detection approach performed better than other GGO detection methods in the experimental results, which achieved a sensitivity of 100% and a specificity of 33.13%.

Image co-segmentation is the problem of extracting common objects from multiple images and it is a very challenging task. In this paper we try to address the co-segmentation problem by embedding image saliency into active contour model. Active contour is a very famous and effective image segmentation method but performs poor results if applied directly to co-segmentation. Therefore, we can introduce additional information to improve the segmentation results, such as saliency which can show the region of interest. In order to optimize the model, we propose an efficient level-set optimization method based on super-pixels, hierarchical computation and convergence judgment. We evaluated the proposed method on iCoseg and MSRC datasets. Compared with other methods, our method yielded better results and demonstrated the significance of using image saliency in active contour.

In computer vision, image retrieval remained a significant problem and recent resurgent of image retrieval also relies on other postprocessing methods to improve the accuracy instead of solely relying on good feature representation. Our method addressed the shape retrieval of binary images. This paper proposes a new integration scheme to best utilize feature representation along with contextual information. For feature representation we used articulation invariant representation; dynamic programming is then utilized for better shape matching followed by manifold learning based postprocessing modified mutual kNN graph to further improve the similarity score. We conducted extensive experiments on widely used MPEG-7 database of shape images by so-called bulls-eye score with and without normalization of modified mutual kNN graph which clearly indicates the importance of normalization. Finally, our method demonstrated better results compared to other methods. We also computed the computational time with another graph transduction method which clearly shows that our method is computationally very fast. Furthermore, to show consistency of postprocessing method, we also performed experiments on challenging ORL and YALE face datasets and improved baseline results.