IntroductionIndian The detection and preventions of criminal activities

IntroductionIndian airline passenger traffic crossed 100 millions in year 2016. The growth was 23.3 percent. The growth in the passenger traffic also imposes security risks as any baggage could be potentially used to smuggle restricted items across the borders. Manually checking the baggages is the simplest security tool. But physically inspecting each container is time consuming. Also to preserve passengers’ privacy, non-intrusive techniques like X-Ray imaging is used to produce detail image of the contents inside the container.X-Ray ImagingX rays are a form of electromagnetic radiations having wavelength ranging from 0.01 nm to 10 nm. The X-Ray beam is transmitted through the object to be scanned. Depending on the thickness and absorption coefficient, some of the X-Rays are absorbed by the material. The attenuated X-Rays reach the line of detectors which measure the intensity in terms of X-Ray counts. These attenuated intensities are plotted to obtain gray scale images.Currently these images are manually inspected by the operators for prohibited items. This is a very challenging visual task as the images often tend to be cluttered. Also there are many objects whose appearance resembles to the target object. With increasing volume of images, inspecting each and every image is painstaking as well as time consuming. The detection and preventions of criminal activities must be done with minimal disruption to the flow of commerce, so it is necessary to automate the X-Ray image inspection process. The efforts are being made to automate classification of X-Ray image into illicit and non-illicit image, so the image analysis framework will detect the images which are likely to be containing prohibited items and a flag will be raised. Objective of the workTo develop an image inspection framework based on modern machine vision and learningtechniques that aims to assist the inspectors by partially automating the inspection process.The aim of the framework is to enable the inspectors to focus the attention on images that arelikely to be anomalous, thus easing the inspection time constraint.Soft ComputingAutomating the X-Ray image classification requires some intelligence to be incorporated with image processing. This can be done using soft computing techniques. Soft computing is an emerging approach to computing which parallel the remarkable ability of the human mind to reason and learn in an environment of uncertainty, imprecision and partial truth. Soft computing is based on some biological inspired methodologies such as genetics, evolution, human nervous systems, etc.(reference: nptel). Soft computing techniques are computational techniques which attempt to analyze very complex phenomena, for which conventional hard computing methods have not yielded low cost, analytic, and complete solutions. (reference: book)  The components of soft computing include Machine learning, fuzzy logic, Evolutionary Computations, Ideas about probability etc.Zadeh (reference?) established the definition of soft computing in the following terms: “Basically, soft computing is not a homogeneous body of concepts and techniques. Rather, it is a partnership of distinct methods that in one way or another conform to its guiding principle. At this juncture, the dominant aim of soft computing is to exploit the tolerance for imprecision and uncertainty to achieve tractability, robustness and low solutions cost. The principal constituents of soft computing are fuzzy logic, neurocomputing, and probabilistic reasoning, with the latter subsuming genetic algorithms, belief networks, chaotic systems, and parts of learning theory. In the partnership of fuzzy logic, neurocomputing, and probabilistic reasoning, fuzzy logic is mainly concerned with imprecision and approximate reasoning; neurocomputing with learning and curve-fitting; and probabilistic reasoning with uncertainty and belief propagation”. Why Soft Computing?The two major problem solving techniques includeHard computing Soft computing Hard Computing deals with precise models where accurate solutions are achieved. Whereas, soft computing deals with approximate models and gives approximate solution to real life complex problems. Complex systems arising in biology, medicine, the humanities, management sciences, and similar fields often remained intractable to conventional mathematical and analytical methods. The two problem solving technologies are shown in the figure.The left diagram shows the traditional hard computing approach where an exact model of the problem is available and traditional mathematical methods are used to solve the problem. The right diagram shows soft computing approach where only an approximate model of the problem may be available, and the solution depends upon approximate reasoning techniques. In threat object recognition task, the uncertainty arises due to variation in background, size, shape and the orientation of the object to be recognised, occlusion etc. It is difficult to find an exact model considering these uncertainties. Soft computing emphasize gains in understanding system behavior in exchange for unnecessary precision. Thus soft computing techniques are used as a practical tool for Image classification based on threat object recognition task.Soft Computing ComponentsMachine LearningArthur Samuel, a pioneer of Artificial Intelligence coined the term machine learning in 1959. Machine learning is a field of computer science that gives computers the ability to learn without being explicitly programmed.(Reference:  Samuel, Arthur (1959). “Some Studies in Machine Learning Using the Game of Checkers”. IBM Journal of Research and Development. 3 (3). doi:10.1147/rd.33.0210..) Learning in this case is defined as ‘A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, increases with experience E.’ (Reference: Mitchell, T. (1997). Machine Learning. McGraw Hill. p. 2. ISBN 0-07-042807-7.)For example, handwritten digits learning problem:Task T: recognizing and classifying handwritten words within images Performance measure P: percent of words correctly classified Training experience E: a database of handwritten words with given classifications In general, any machine learning problem can be assigned to one of two broad classifications: Supervised learning, Unsupervised learning and Reinforcement learning.Supervised Learning:Classification Problems:  Inputs are divided into two or more classes. For example, to predict whether the tumor is malignant or benign.Algorithms for classificationDecision TreesLogistic RegressionNaive BayesK Nearest NeighborsLinear SVC (Support vector Classifier)Regression Problems: The outputs are continuous rather than discrete. For example, Given a size of the house predict the price(real value).Regression AlgorithmsLinear RegressionRegression Trees(e.g. Random Forest)Support Vector Regression (SVR)Depending on the nature of interaction between learner and the environment, machine learning algorithms can be classified as follows:Supervised Learning: Supervised learning is the machine learning task of learning relationship between input and output from labeled training data. The training data consist of a set of training examples. In supervised learning, each example is a pair consisting of an input object (feature vector) and a desired output value. A supervised learning algorithm analyzes the training data and produces an inferred function, which can be used for mapping new examples. This requires the learning algorithm to generalize from the training data to unseen situations. (reference: http://dataaspirant.com/2014/09/19/supervised-and-unsupervised-learning/)Supervised learning problems are categorized into Regression and Classification problems. Regression: The results of regression are within a continuous output, The algorithm tries to map input variables to some continuous function. The input feature set in training data has real valued labels.Classification: In classification problem, the outputs are continuous rather than discrete. The labels in the training set are one of the discrete categories to which the input features belong.Unsupervised Learning: Unsupervised machine learning is drawing inferences from data which does not have labels associated with it. Since, unsupervised learning is the task of deriving structure from a data where the effect of its variables is unknown, so unlike Supervised or reinforcement learning, there is no feedback based on prediction results. For example, Clustering. Clustering refers to grouping observations together in such a way that members of a common group are similar to each other, and different from members of other groups. A common application is to identify segments of customers or prospects with similar preferences or buying habits. Reinforcement Learning: In reinforcement learning, unlike supervised learning, machine is not provided with examples of correct input-output pairs, a method for the machine to quantify its performance in the form of a reward signal. Reinforcement learning methods resemble how humans and animals learn: the machine tries a bunch of different things and is rewarded when it does something well.Fuzzy LogicFuzzy logic is a form of many-valued logic in which the truth values of variables may be any real number between 0 and 1. It is employed to handle the concept of partial truth, where the truth value may range between completely true and completely false. By contrast, in Boolean logic, the truth values of variables may only be the integer values 0 or 1.A fuzzy set is a set without a crisp, clearly defined boundary. It can contain elements with only a partial degree of membership. The degree an object belongs to a fuzzy set is denoted by a membership value between 0 and 1. For example, the girl is tall to the degree 0.8. A membership function associated with a given fuzzy set maps an input value to its appropriate membership value.Fuzzy Logic incorporates a simple, rule-based IF X AND Y THEN Z approach to a solving control problem rather than attempting to model a system mathematically. The Fuzzy Logic model is empirically-based, relying on an operator’s experience rather than their technical understanding of the system. For example, rather than dealing with temperature control in terms such as “SP =500F”, “T <1000F", or "210C