Narrow Results

The pivotal task of remote sensing image (RSI) processing change detection (CD) highly aims to accurately detect changes in land cover based on multi-temporal images. With the advent of deep learning, technology has delivered remarkable results in the last years in the detection of variations in forest land cover data. Some of the conventional CD techniques are weak and are highly susceptible to errors and can result even in inaccurate outcomes. Thus, certain techniques are not desirable for real-time CD applications. To abridge this gap, this research introduces an innovative work for forest CD utilizing the proposed Taylor Shepherd Golden Optimization_ResUNet (TSGO_ResUNet) and Fuzzy Neural network (Fuzzy NN) for segment mapping. Here, the segmentation process is accomplished using ResUNet to determine the exact boundary or shape of each object for every pixel in the image. Furthermore, TSGO is achieved by consolidating Taylor Shuffled Shepherd Optimization (TSSO) with Golden Search Optimization (GSO). In addition, the devised TSGO_ResUNet + Fuzzy NN has gained maximum accuracy and kappa coefficient of 0.952 and 0.785, and minimum error rate of 0.051.

A nonparametric unsupervised method for analyzing changes in complex datasets is proposed. It is based on the mean shift clustering algorithm. Mean shift is used to cluster the old and new datasets and compare the results in a nonparametric manner. Each point from the new dataset naturally belongs to a cluster of points from its dataset. The method is also able to find to which cluster the point belongs in the old dataset and use this information to report qualitative differences between that dataset and the new one. Changes in local cluster distribution are also reported. The report can then be used to try to understand the underlying reasons which caused the changes in the distributions. On the basis of this method, a transductive transfer learning method for automatically labeling data from the new dataset is also proposed. This labeled data is used, in addition to the old training set, to train a classifier better suited to the new dataset. The algorithm has been implemented and tested on simulated and real (a stereo image pair) datasets. Its performance was also compared with several state-of-the-art methods.

Many investigations of image sequences can be understood on the basis of a few concepts for which computational approaches become increasingly available. The estimation of optical flow fields is discussed, exhibiting a common foundation for feature-based and differential approaches. The interpretation of optical flow fields is mostly concerned so far with approaches which infer the 3-D structure of a rigid point configuration in 3-D space and its relative motion with respect to the image sensor from an image sequence. The combination of stereo and motion provides additional incentives to evaluate image sequences, especially for the control of robots and autonomous vehicles. Advances in all these areas lead to the desire to describe the spatio-temporal development recorded by an image sequence not only at the level of geometry, but also at higher conceptual levels, for example by natural language descriptions.

Historically, data mining research has been focused on discovering sets of attributes that discriminate data entities into classes or association rules between attributes. In contrast, we are working to develop data mining techniques to discover patterns consisting of complex relationships between entities. Our research is particularly applicable to domains in which the data is event driven, such as counter-terrorism intelligence analysis. In this paper we describe an algorithm designed to operate over relational data received from a continuous stream. Our approach includes a mechanism for summarizing discoveries from previous data increments so that the globally best patterns can be computed by examining only the new data increment. We then describe a method by which relational dependencies that span across temporal increment boundaries can be efficiently resolved so that additional pattern instances, which do not reside entirely in a single data increment, can be discovered. We also describe a method for change detection using a measure of central tendency designed for graph data. We contrast two formulations of the change detection process and demonstrate the ability to identify salient changes along meaningful dimensions and recognize trends in a relational data stream.

Characterizing users and tracking their activities in online classified ads is a topic of great importance. However, some of the underlying problems associated with modeling users and detecting their behavioral changes have not been well-studied.

In this paper, we develop a probabilistic framework for characterizing users and quantifying some of the spatial and temporal variations in their posts. Our work on characterizing users study the problem in the context of detecting if a user belongs to a class, based on the ads the user has posted. Our approach is based on user profiling, where given statistics on user posts, the affinity of a user to a class is estimated. We show how profiles can be constructed with and without training data and report the effectiveness of our approaches in detecting two user classes business and non-business.

Our work on quantifying changes due to spatial and temporal variations is based on a probabilistic model of user behavior and a generative model that can predict ad posts from each location. We evaluate these models on a relatively large set of users and ads, and report our results on two classes of users monitored over a period of almost a year.

In this paper a novel technique for building change detection from remote sensing imagery is presented. It includes two main stages: (1) Object-specific discriminative features are extracted using Morphological Building Index (MBI) to automatically detect the existence of buildings in remote sensing images. (2) Pixel-based image matching is measured on the basis of Mutual Information (MI) of the images by Normalized Mutual Information (NMI). Here, the MBI features values are computed for each of the pair images taken over the same region at two different times and then changes in these two MBI images are measured to indicate the building change. MI is estimated locally for all the pixels for image matching and then thresholding is applied for eliminating those pixels which are responsible for strong similarity. Finally, after getting the MBI and NMI images, a further fusion of these two images is done for refinement of the change result. For evaluation purpose, the experiments are carried on QuickBird, IKONOS images and images taken from Google Earth. The results show that the proposed technique can attain acceptable correctness rates above 90% with Overall Accuracy (OA) 89.52%.

Online clustering has attracted attention due to the explosion of ubiquitous continuous sensing. Streaming clustering algorithms need to look for new structures and adapt as the data evolves, such that outliers are detected, and that new emerging clusters are automatically formed. The performance of a streaming clustering algorithm needs to be monitored over time to understand the behavior of the streaming data in terms of new emerging clusters and number of outlier data points. Small datasets with 2 or 3 dimensions can be monitored by plotting the clustering results as data evolves. However, as the size and dimensions of streaming data increase, plotting the clustering result becomes unfeasible. Therefore, incremental internal Validity Indices (iCVIs) could be applied for monitoring the performance of an online clustering algorithm. In this paper, we study the internal incremental Davies-Bouldin (iDB) cluster validity index in the context of big streaming data analysis. Also, we study the effect of large number of samples on the values of the iCVI (iDB). Finally, we propose a way to project streaming data into a lower space for cases where the distance measure does not perform as expected in the high dimensional space.

In the management of large enterprise communication networks, it becomes difficult to detect and identify causes of abnormal change in traffic distributions when the underlying logical topology is dynamic. This paper describes a novel approach to abnormal network change detection by representing periodic observations of logical communications within a network as a time series of graphs. A number of graph distance measures are proposed to assess the difference between successive graphs and identify abnormal behaviour. Localisation techniques have also been described to show where in the network most change occurred.

This paper introduces a graphical user interface approach to facilitate an efficient and timely generation of statistic data from input videos. By means of a carefully-designed graphical user interface, users can interactively add in various kinds of markers, known as the statistic inducers, on the screen of an input video to specify the areas of interest corresponding to the locations of relevant events. These inducers are in the form of two-dimensional points, lines, polygons, and grids, and can be put on the video screen with great ease. Using these inducers, we not only can efficiently customize the system for a given statistic generation task; in addition, we can also precisely constrain the time-consuming space-time video analysis process (as well as any additional analysis process like optical flow computation or object recognition) on the user-specified areas. To demonstrate the efficacy of the proposed approach, we developed a prototypic system and experimented it in two different statistic generation cases: dormitory light switching and road traffic. In both cases, we just need a few minutes of UI customization time to set up the inducers; once this is done, timely statistics can be automatically generated subsequently.

Nowadays, hyperspectral imaging (HSI) attracts the interest of many researchers in solving the remote sensing problems especially in various specific domains such as agriculture, snow/ice, object detection and environmental monitoring. In the previous literature, various attempts have been made to extract the critical information through hyperspectral imaging which is not possible through multispectral imaging (MSI). The classification in image processing is one of the important steps to categorize and label the pixels based on some specific rules. There are various supervised and unsupervised approaches which can be used for classification. Since the past decades, various classifiers have been developed and improved to meet the requirement of remote sensing researchers. However, each method has its own merits and demerits and is not applicable in all scenarios. Past literature also concluded that deep learning classifiers are more preferable as compared to machine learning classifiers due to various advantages such as lesser training time for model generation, handle complex data and lesser user intervention requirements. This paper aims to perform the review on various machine learning and deep learning-based classifiers for HSI classification along with challenges and remedial solution of deep learning with hyperspectral imaging. This work also highlights the various limitations of the classifiers which can be resolved with developments and incorporation of well-defined techniques.

Data stream mining has attracted considerable attention over the past few years owing to the significance of its applications. Streaming data is often evolving over time. Capturing changes could be used for detecting an event or a phenomenon in various applications. Weather conditions, economical changes, astronomical, and scientific phenomena are among a wide range of applications. Because of the high volume and speed of data streams, it is computationally hard to capture these changes from raw data in real-time. In this paper, we propose a novel algorithm that we term as STREAM-DETECT to capture these changes in data stream distribution and/or domain using clustering result deviation. STREAM-DETECT is followed by a process of offline classification CHANGE-CLASS. This classification is concerned with the association of the history of change characteristics with the observed event or phenomenon. Experimental results show the efficiency of the proposed framework in both detecting the changes and classification accuracy.

Several studies have focused on mining changes in different time-period databases. Analyzing these change behaviors provides useful information for managers to develop better marketing strategies and decision making. Although some researchers have developed efficient methods for association rule change detection, no attempt has been made to analyze time-interval sequential pattern changes in databases collected over time. Therefore, this research proposes a time-interval sequential pattern change detection framework to derive the change trends in customer behaviors in two periods. First, two time-interval sequential pattern sets are generated from two time-period databases respectively using the proposed DTI-Apriori algorithm. Different from previous mining methods that require users to manually define a set of time-interval ranges in advance, the DTI-Apriori algorithm automatically arranges the time-interval range and then generates time-interval sequential patterns. The degree of change for each pair of time-interval sequential patterns from different time periods is evaluated next. Based on the degree of change, a time-interval sequential pattern is clarified as one of the following three change types: an emerging time-interval sequential pattern, an unexpected time-interval sequential pattern, or an added/perished time-interval sequential pattern. Significant change patterns are returned to users for further analysis if the degree of change is large enough.

Faces are socially very important visual objects and the detection of a change in faces is an essential evolutionary skill. To investigate whether configural computation of faces automatically occurs under non-attentional condition, visual mismatch negativity (vMMN) elicited by deviant orientation (90° vs. 0°) of faces was analyzed using the equi-probable paradigm which eliminated the low-level refractory effects. Fourteen participants were tested and schematic face stimuli were used. In comparison with control face stimuli, the deviant orientation of faces elicited larger N170 and smaller P2. During the time range between 100–300 ms post stimulus onset, face orientation changes elicited occipital–temporal distributed vMMN. The source analysis of face-MMN showed that it was generated in both temporal and frontal lobes. These data supported the hypotheses that the disruption of facial configuration processing caused by inverted faces is relatively independent of attentional resources.

As the contamination over the surface of the earth is increasing exponentially, the land cover and land use detection techniques are considered as important elements in mapping and monitoring the land degradation. Remote sensing plays a vital role in identifying the land changes over the period of time. As land degradation occurs, resource demand will increase and reliable service to achieve land neutrality will increase. Connected device (IoT) could be used to achieve this neutrality in an intelligent and effective manner. Innumerable change detection methods have been developed for as far back as five decades. These studies deal in detail about the different satellite imagery data, image preprocessing techniques and the discussion of pixel-based and object-based change detection techniques. In addition, the dataset, preprocessing and change detection technique are interrelated with each other and their connection between the techniques are clarified dependent on the element of image analysis. The merits and limitation of different methods are also explained in detail.

Near real-time monitoring of abrupt changes in satellite time series is important for timely warning of land covers changes. Regression model-based method has been frequently used to detect abrupt change (outlier or anomaly) in time series data. Abrupt change is often determined by residuals test after regression. A simple and widely used residuals test technique is confidence interval (CI), which is often time-independent or constant in many studies. However, satellite time series data is characterized by seasonal variability and periodicity. Although the periodicity could be fitted well by a seasonal-trend regression model, the seasonal variability still remains in the residuals of the regression model. The seasonal variability would lead to less reliable results if abrupt changes are detected by a constant confidence interval (CCI). In order to improve the reliability of abrupt change monitoring in satellite time series, in this paper we develop a criterion namely seasonal confidence interval (SCI) of regression residuals. Experimental evaluations with some simulated and actual satellite time series data demonstrate better performance of the proposed SCI criterion than the CCI criterion for monitoring abrupt changes in satellite time series.

Background Subtraction (BgS) is a widely researched technique to develop online Change Detection algorithms for static video cameras. Many BgS methods have employed the unsupervised, adaptive approach of Gaussian Mixture Model (GMM) to produce decent backgrounds, but they lack proper consideration of scene semantics to produce better foregrounds. On the other hand, with considerable computational expenses, BgS with Deep Neural Networks (DNN) is able to produce accurate background and foreground segments. In our research, we blend both approaches for the best. First, we formulated a network called Convolutional Density Approximation (CDA) for direct density estimation of background models. Then, we propose a self-supervised training strategy for CDA to adaptively capture high-frequency color distributions for the corresponding backgrounds. Finally, we show that background models can indeed assist foreground extraction by an efficient Neural Motion Subtraction (NeMos) network. Our experiments verify competitive results in the balance between effectiveness and efficiency.

Most classification methods are based on the assumption that the historic data involved in building and verifying the model is the best estimator of what will happen in the future. One important factor that must not be set aside is the time factor. As more data is accumulated into the problem domain, incrementally over time, one must examine whether the new data agrees with the previous datasets and make the relevant assumptions about the future. This work presents a new change detection methodology, with a set of statistical estimators. These changes can be detected independently of the data mining algorithm, which is used for constructing the corresponding model. By implementing the novel approach on a set of artificially generated datasets, all significant changes were detected in the relevant periods. Also, in the real-world datasets evaluation, the method produced similar results.

Ants, bees and other social insects deposit pheromone (a type of chemical) in order to communicate between the members of their community. Pheromone that causes clumping or clustering behavior in a species and brings individuals into a closer proximity is called aggregation pheromone. This article presents a novel method for change detection in remotely sensed images considering the aggregation behavior of ants. Change detection is viewed as a segmentation problem where changed and unchanged regions are segmented out via clustering. At each location of data point, representing a pixel, an ant is placed; and the ants are allowed to move in the search space to find out the points with higher pheromone density. The movement of an ant is governed by the amount of pheromone deposited at different points of the search space. More the deposited pheromone, more is the aggregation of ants. This leads to the formation of homogenous groups of data. Evaluation on two multitemporal remote sensing images establishes the effectiveness of the proposed algorithm over an existing thresholding algorithm.