The remote sensing community has developed algorithms to extract diverse geographic information from electro-optical and radar imagery, and study them along with data from other weather sensors such as wind, temperature and humidity. Simultaneously, the computer vision community has adopted machine learning algorithms for detection problems with promising success. Many such algorithms utilize labeled datasets to learn object signatures for remote sensing. Meanwhile, research in the data mining domain also has evolved spatial algorithms for discovering association and outliers in the geographic data. In a first, the current work introduces techniques developed in data mining domain to the computer vision community utilizing object detection datasets for applications in remote sensing. Specifically, this work applies the concepts of rule mining to geo-registered aerial object detection datasets.Aerial object detection datasets were originally intended for training a classifier. However, this work innovates the use of ground-truth labels with geo-location to identify co-located classes. The paper establishes a method to adapt the aerial object detection dataset to mining friendly data. A theoretical analysis of the co-location frequency and size of search area is presented. The search area size for a set of objects is heuristically estimated. The work utilizes state-of-the-art data mining methods and uses the Support and Lift metrics to study the effects of varying count of object instances per class. In addition to finding prominent co-locations and anomalies, this work demonstrates how composite objects can be detected, which have varying spatial configurations of their constituents. The developed method is tested on publicly available XView (60-classes) and FAIR1M (37-classes) datasets.

The advanced malware continue to be a challenge in digital world that signature-based detection techniques fail to conquer. The malware use many anti-detection techniques to mutate. Thus no virus scanner can claim complete malware detection even for known malware. Static and dynamic analysis techniques focus upon different kinds of malware such as Evasive or Metamorphic malware. This paper proposes a comprehensive approach that combines static checking and dynamic analysis for malware detection. Static analysis is used to check the specific code characteristics. Dynamic analysis is used to analyze the runtime behavior of malware. The authors propose a framework for the automated analysis of an executable's behavior using text mining. Text mining of dynamic attributes identifies the important features for classifying the executable as benign and malware. The synergistic combination proposed in this paper allows detection of not only known variants of malware but even the obfuscated, packed and unknown malware variants and malware evasive to dynamic analysis.

Content Based Retrieval (CBR) systems use Relevance Feedback (RF) to fill the semantic gap. RF can be short-term or long-term. The introduction of long-term learning methods address the memory problem in short-term learning methods. In this letter we propose a new method to enhance the gain of long-term relevance feedback. We have come up with a long term learning scheme in relevance feedback for CBR. The proposed system integrates the user feedback from all iterationations and instills memory into the feedback system of CBR without saving any log of earlier retrievals. In this letter, we have come up with a method to update the cluster parameters and weights assigned to features by accumulating the knowledge obtained from the user over iterations. The proposed update method is validated in the image retrieval context in terms of conventional recall-precision graph and retrieval accuracy.

This paper presents feature selection algorithms for multilayer perceptrons (MLPs) and multiclass support vector machines (SVMs), using mutual information between class labels and classifier outputs, as an objective function. This objective function involves inexpensive computation of information measures only on discrete variables; provides immunity to prior class probabilities; and brackets the probability of error of the classifier. The maximum output information (MOI) algorithms employ this function for feature subset selection by greedy elimination and directed search. The output of the MOI algorithms is a feature subset of user-defined size and an associated trained classifier (MLP/SVM). These algorithms compare favorably with a number of other methods in terms of performance on various artificial and real-world data sets.

The Laplacian pyramid was proposed as a compact image code by Burt (1983). However, it is shown that it does not achieve optimal signal compaction due to significant overlap between subbands in the frequency domain. In addition, it also leads to a slightly redundant signal representation. An extension of the Laplacian pyramid is proposed that exploits this redundancy to achieve better signal compaction. In contrast to the unidirectional fine to coarse approach used by all multiscale representations, the proposed signal representation adopts a bi-directional approach. Each subband finally depends on both its coarser and finer neighboring subbands. The representation is symmetric in the sense that each subband is a product of an approximation process (based on a finer approximation) as well as a prediction process (based on a coarser approximation). Applications are in the areas of image characterisation, zooming and compression.

Identifying relevant features for a classification task is an important issue in machine learning.In this paper, we present a feature crediting scheme for multiclass pattern recognition tasks, that utilizes the ability of Support Vector Machines to generalize well in high dimensional feature spaces.Support Vector learning identifies a small subset of training data relevant for the classification task.They primarily tackle the binary classification problem.This scheme uses relevant examples to identify relevant features for multi-class classification.We present, and employ for this purpose, an informationtheoretic measure of classifier performance.This measure addresses the key issue of average rate of information being delivered by the classifier.It provides immunity to sampling bias in the data and sensitivity to pattern of errors made by the classifier.Empirical results on a number of datasets suggest efficient applicability to data with a very large number of features.

The issues governing the computation of optical flow in image sequences are addressed. The trade-off between accuracy versus computation cost is shown to be dependent on the redundancy of the image representation. This dependency is highlighted by reformulating Horn's (1986) algorithm, making explicit use of the approximations to the continuous basis functions underlying the discrete representation. The computation cost of estimating optical flow, for a fixed error tolerance, is shown to be a minimum for images resampled at twice the Nyquist rate. The issues of derivative calculation and multiresolution representation are also briefly discussed in terms of basis functions and information encoding. A multiresolution basis function formulation of Horn's algorithm is shown to lead to large improvements in dealing with high frequencies and large displacements.

The study of multiresolution representations of images led to the development of the pyramid and wavelet representations. The redundancy in the pyramid representation bestows on it the properties of an error-correcting code. The error-correcting properties of the pyramids studied are a property of the representation itself and not a result of making any assumptions about the original images. The article extends the error-reduction properties of frames to an error-correcting routine. It leads to complete noise removal for sparse noise for a class of frames. The results for images using Laplacian pyramids show an additional 10-dB improvement over the previously demonstrated noise reduction for frames.

A given (overcomplete) discrete oriented pyramid may be converted into a steerable pyramid by interpolation. We present a technique for deriving the optimal interpolation functions (otherwise called 'steering coefficients'). The proposed scheme is demonstrated on a computationally efficient oriented pyramid, which is a variation on the Burt and Adelson (1983) pyramid. We apply the generated steerable pyramid to orientation-invariant texture analysis in order to demonstrate its excellent rotational isotropy. High classification rates and precise rotation identification are demonstrated.>