Search

This survey presents a comprehensive overview of machine learning methods for cybersecurity intrusion detection systems, with a specific focus on recent approaches based on deep learning (DL). The review analyzes recent methods with respect to their intrusion detection mechanisms, performance results, and limitations as well as whether they use benchmark databases to ensure a fair evaluation. In addition, a detailed investigation of benchmark datasets for cybersecurity is presented. This article is intended to provide a road map for readers who would like to understand the potential of DL methods for cybersecurity and intrusion detection systems, along with a detailed analysis of the benchmark datasets used in the literature to train DL models.

Explainable autonomous driving systems (EADS) are emerging recently as a combination of explainable artificial intelligence (XAI) and vehicular automation (VA). EADS explains events, ambient environments, and engine operations of an autonomous driving vehicular, and it also delivers explainable results in an orderly manner. Explainable semantic segmentation (ESS) plays an essential role in building EADS, where it offers visual attention that helps the drivers to be aware of the ambient objects irrespective if they are roads, pedestrians, animals, or other objects. In this paper, we propose the first ESS model for EADS based on the variational autoencoder (VAE), and it uses the multiscale second-order derivatives between the latent space and the encoder layers to capture the curvatures of the neurons’ responses. Our model is termed as Mgrad 2 VAE and is bench-marked on the SYNTHIA and A2D2 datasets, where it outperforms the recent models in terms of image segmentation metrics.

Touchless palmprint recognition systems enable high-accuracy recognition of individuals through less-constrained and highly usable procedures that do not require the contact of the palm with a surface. To perform this recognition, methods based on local texture descriptors and convolutional neural networks (CNNs) are currently used to extract highly discriminative features while compensating for variations in scale, rotation, and illumination in biometric samples. In particular, the main advantage of CNN-based methods is their ability to adapt to biometric samples captured with heterogeneous devices. However, the current methods rely on either supervised training algorithms, which require class labels (e.g., the identities of the individuals) during the training phase, or filters pretrained on general-purpose databases, which may not be specifically suitable for palmprint data. To achieve a high-recognition accuracy with touchless palmprint samples captured using different devices while neither requiring class labels for training nor using pretrained filters, we introduce PalmNet, which is a novel CNN that uses a newly developed method to tune palmprint-specific filters through an unsupervised procedure based on Gabor responses and principal component analysis (PCA), not requiring class labels during training. PalmNet is a new method of applying Gabor filters in a CNN and is designed to extract highly discriminative palmprint-specific descriptors and to adapt to heterogeneous databases. We validated the innovative PalmNet on several palmprint databases captured using different touchless acquisition procedures and heterogeneous devices, and in all cases, a recognition accuracy greater than that of the current methods in this paper was obtained.

Image-based methods for estimating the particle size distribution (granulometry) usually analyze two-dimensional (2-D) samples of particles disposed on a conveyor belt. Such approaches have to deal with occlusions and cannot evaluate the thickness of each particle. Three-dimensional (3-D) vision systems can reduce the acquisition constraints and speed up the quality control process. This paper proposes a novel 3-D vision system for analyzing the granulometry of falling particles. The system is designed to work in real time and to compute a partial 3-D reconstruction of the particle from a single pair of two-view images, which is then enhanced by using a neural-based technique. The validation of the proposed approach has been performed by considering three application scenarios for which the system achieved satisfactory accuracy and robustness.

Computer Aided Diagnosis (CAD) systems are increasingly utilizing image analysis and Deep Learning (DL) techniques, due to their high accuracy in several medical imaging fields, including the detection of Acute Lymphoblastic (or Lymphocytic) Leukemia (ALL) from peripheral blood samples. However, no method in the literature has specifically analyzed the focus quality of ALL images or proposed a technique for sharpening the samples in an adaptive way for the purpose of classification. To address this issue, in this paper we propose the first machine learning-based approach able to enhance blood sample images by an adaptive unsharpening method. The method uses image processing techniques and DL to normalize the radius of the cell, estimate the focus quality, adaptively improve the sharpness of the images, and then perform the classification. We evaluated the methodology on a public database of ALL images, considering several state-of-the-art CNNs to perform the classification, with results showing the validity of the proposed approach. For a complete reproducibility of the work, the source code is available at: http://iebil.di.unimi.it/cnnALL/index.htm.

Automotive companies are investing a relevant amount of resources for designing autonomous driving systems, driver assistance technologies, and systems for assessing the driver’s attention. In this context, two important applications consist in processing images of the surrounding environment to respectively separate the different objects in the scene (semantic segmentation) and to estimate their distances. In both applications, methods based on Deep Learning (DL) and Convolutional Neural Networks (CNN) are being increasingly used, considering LiDAR scans or RGB images. However, LiDAR scanners require dedicated sensors, high costs, and post-processing algorithms to estimate a dense depth map or a three-dimensional representation of the surrounding environment. Moreover, current methods in the literature based on RGB images do not consider the combination of semantic segmentation and depth estimation for assessing the distances of specific objects in the scene. In this paper, we propose the first method in the literature able to estimate the distances of pedestrians/cyclists from the vehicle by using only an RGB image and CNNs, without the need for any LiDAR scanner or any device designed for the three-dimensional reconstruction of the scene. We evaluated our approach on a public dataset of RGB images captured in an automotive scenario, with results confirming the feasibility of the proposed method.

Unsupervised Learning (UL) models are a class of Machine Learning (ML) which concerns with reducing dimensionality, data factorization, disentangling and learning the representations among the data. The UL models gain their popularity due to their abilities to learn without any predefined label, and they are able to reduce the noise and redundancy among the data samples. However, generalizing the UL models for different applications including image generation, compression, encoding, and recognition faces different challenges due to limited available data for learning, diversity, and complex dimensions. To overcome such challenges, we propose a partial learning procedure by utilizing the $\beta$-Non Negative Matrix Factorization ($\beta$-NMF), which maps the data into two complementary subspaces constituting generalized driven priors among the data. Moreover, we employ a dual-shallow Autoencoder (AE) to learn the subspaces separately or jointly for image reconstruction and visualization tasks, where our model performance shows superior results to the literary works when learning the model with a small amount of data and generalizing it for large-scale unseen data.

Unsupervised Learning (UL) methods are a class of machine learning which aims to disentangle the representations and reduce the dimensionality among the data without any predefined labels. Among all UL methods, the Non-negative Matrix Factorization (NMF) factorizes the data into two subspaces of non-negative components. Moreover, the NMF enforces the non-negativity, sparsity, and part-based analysis, thus the representations can be interpreted and explained for the Explainable Artificial Intelligence (XAI) applications. However, one of the main issues when using the NMF is to impose the factorization rank r to identify the dimensionality of the subspaces, where the rank is usually unknown in advance and known as the non-negative rank that is used as a prior to carrying out the factorization. Accordingly, we propose a novel method for the non-negative rank r approximation to help solving this problem, and we generalize our method among different image scales. Where, the results on different image data sets confirm the validity of our approach.

Generative Adversarial Networks (GAN) are being increasingly used to perform face aging due to their capabilities of automatically generating highly-realistic synthetic images by using an adversarial model often based on Convolutional Neural Networks (CNN). However, GANs currently represent black box models since it is not known how the CNNs store and process the information learned from data. In this paper, we propose the first method that deals with explaining GANs, by introducing a novel qualitative and quantitative analysis of the inner structure of the model. Similarly to analyzing the common genes in two DNA sequences, we analyze the common filters in two CNNs. We show that the GANs for face aging partially share their parameters with GANs trained for heterogeneous applications and that the aging transformation can be learned using general purpose image databases and a fine-tuning step. Results on public databases confirm the validity of our approach, also enabling future studies on similar models.

Biometric systems based on touchless and less-constrained palmprint are being increasingly studied since they allow a favorable trade-off between high-accuracy and high usability recognition. Another advantage is that with a palmar hand acquisition, it is possible to extract the palmprint as well as the Inner Finger Texture (IFT) and increase the recognition accuracy without requiring further biometric acquisitions. Recently, most methods in the literature consider Deep Learning (DL) and Convolutional Neural Networks (CNN), due to their high recognition accuracy and the capability to adapt to biometric samples captured in heterogeneous and less-constrained conditions. However, current methods based on DL do not consider the fusion of palmprint with IFT. In this work, we propose the first novel method in the literature based on a CNN to perform the fusion of palmprint and IFT using a single hand acquisition. Our approach uses an innovative procedure based on training the same CNN topology separately on the palmprint and the IFT, adapting the neural model to the different biometric traits, and then performing a feature-level fusion. We validated the proposed methodology on a public database captured in touchless and less-constrained conditions, with results showing that the fusion enabled to increase the recognition accuracy, without requiring multlple biometric acquisltions.

Cloud Computing (CC) is a model that enables ubiquitous, convenient, and on-demand network access to a shared pool of configurable computing resources. In CC applications, it is possible to access both software and hardware architectures remotely and with little or no knowledge about their physical or logical locations. Due to its low deployment and management costs, the CC paradigm is being increasingly used in a wide variety of online services and applications, including remote computation, software-as-a-service, off-site storage, entertainment, and communication platforms. However, several aspects of CC applications, such as system design, optimization, and security issues, have become too complex to be efficiently treated using traditional algorithmic approaches under the increasingly high complexity and performance demands of current applications. Recently, advances in Computational Intelligence (CI) techniques have fostered the development of intelligent solutions for CC applications. CI methods such as artificial neural networks, deep learning, fuzzy logic, and evolutionary algorithms have enabled improving CC paradigms through their capabilities of extracting knowledge from high quantities of real-world data, thus further optimizing their design, performance, and security with respect to traditional techniques. This chapter introduces recent CI techniques, reviews the main applications of CI in CC, and presents challenges and research trends.

Biometric systems consist of devices, procedures, and algorithms used to recognize people based on their physiological or behavioral features, known as biometric traits. Computational intelligence (CI) approaches are widely adopted in establishing identity based on biometrics and also to overcome non-idealities typically present in the samples. Typical areas include sample enhancement, feature extraction, classification, indexing, fusion, normalization, and anti-spoofing. In this context, computational intelligence plays an important role in performing of complex non-linear computations by creating models from the training data. These approaches are based on supervised as well as unsupervised training techniques. This work presents computational intelligence techniques applied to biometrics, from both a theoretical and an application point of view.

Touchless fingerprint recognition systems do not require contact of the finger with any acquisition surface and thus provide an increased level of hygiene, usability, and user acceptability of fingerprint-based biometric technologies. The most accurate touchless approaches compute 3-D models of the fingertip. However, a relevant drawback of these systems is that they usually require constrained and highly cooperative acquisition methods. We present a novel, fully touchless fingerprint recognition system based on the computation of 3-D models. It adopts an innovative and less-constrained acquisition setup compared with other previously reported 3-D systems, does not require contact with any surface or a finger placement guide, and simultaneously captures multiple images while the finger is moving. To compensate for possible differences in finger placement, we propose novel algorithms for computing 3-D models of the shape of a finger. Moreover, we present a new matching strategy based on the computation of multiple touch-compatible images. We evaluated different aspects of the biometric system: acceptability, usability, recognition performance, robustness to environmental conditions and finger misplacements, and compatibility and interoperability with touch-based technologies. The proposed system proved to be more acceptable and usable than touch-based techniques. Moreover, the system displayed satisfactory accuracy, achieving an equal error rate of 0.06% on a dataset of 2368 samples acquired in a single session and 0.22% on a dataset of 2368 samples acquired over the course of one year. The system was also robust to environmental conditions and to a wide range of finger rotations. The compatibility and interoperability with touch-based technologies was greater or comparable to those reported in public tests using commercial touchless devices.

Age estimation based on face images plays an important role in a wide range of scenarios, including security and defense applications, border control, human-machine interaction in ambient intelligence applications, and recognition based on soft biometric information. Recent methods based on deep learning have shown promising performance in this field. Most of these methods use deep networks specifically designed and trained to cope with this problem. There are also some studies that focus on applying deep networks pre-trained for face recognition, which perform a fine-tuning to achieve accurate results. Differently, in this paper, we propose a preliminary study on increasing the performance of pre-trained deep networks by applying postprocessing strategies. The main advantage with respect to fine-tuning strategies consists of the simplicity and low computational cost of the post-processing step. To the best of our knowledge, this paper is the first study on age estimation that proposes the use of post-processing strategies for features extracted using pre-trained deep networks. Our method exploits a set of pre-trained Convolutional Neural Networks (CNNs) to extract features from the input face image. The method then performs a feature level fusion, reduces the dimensionality of the feature space, and estimates the age of the individual by using a Feed-Forward Neural Network (FFNN). We evaluated the performance of our method on a public dataset (Adience Benchmark of Unfiltered Faces for Gender and Age Classification) and on a dataset of nonideal samples affected by controlled rotations, which we collected in our laboratory. Our age estimation method obtained better or comparable results with respect to state-of-the-art techniques and achieved satisfactory performance in non-ideal conditions. Results also showed that CNNs trained on general datasets can obtain satisfactory accuracy for different types of validation images, also without applying fine-tuning methods.

An early wildfire detection is essential in order to assess an effective response to emergencies and damages. In this paper, we propose a low-cost approach based on image processing and computational intelligence techniques, capable to adapt and identify wildfire smoke from heterogeneous sequences taken from a long distance. Since the collection of frame sequences can be difficult and expensive, we propose a virtual environment, based on a cellular model, for the computation of synthetic wildfire smoke sequences. The proposed detection method is tested on both real and simulated frame sequences. The results show that the proposed approach obtains accurate results.

Most fingerprint recognition systems use Level 1 characteristics (ridge flow, orientation, and frequency) and Level 2 features (minutiae points) to recognize individuals. Level 3 features (sweat pores, incipient ridges and ultra-thin characteristics of the ridges) are less frequently adopted because they can be extracted only from high resolution images, but they have the potential of improving all the steps of the biometric recognition process. In particular, sweat pores can be used for quality assessment, liveness detection, biometric matching in live applications, and matching of partial latent fingerprints in forensic applications. Currently, each type of fingerprint acquisition technique (touch-based, touchless, or latent) requires a different algorithm for pore extraction. In this paper, we propose the first method in the literature able to extract the coordinates of the pores from touch-based, touchless, and latent fingerprint images. Our method uses specifically designed and trained Convolutional Neural Networks (CNN) to estimate and refine the centroid of each pore. Results show that our method is feasible and achieved satisfactory accuracy for all the types of evaluated images, with a better performance with respect to the compared state-of-the-art methods.

Automated Border Control (ABC) systems are being increasingly used to perform a fast, accurate, and reliable verification of the travelers' identity. These systems use biometric technologies to verify the identity of the person crossing the border. In this context, fingerprint verification systems are widely adopted due to their high accuracy and user acceptance. Matching score normalization methods can improve the performance of fingerprint recognition in ABC systems and mitigate the effect of non-idealities typical of this scenario without modifying the existing biometric technologies. However, privacy protection regulations restrict the use of biometric data captured in ABC systems and can compromise the applicability of these techniques. Cohort score normalization methods based only on impostor scores provide a suitable solution, due to their limited use of sensible data and to their promising performance. In this paper, we propose a privacy-compliant and adaptive normalization approach for enhancing fingerprint recognition in ABC systems. The proposed approach computes cohort scores from an external public dataset and uses computational intelligence to learn and improve the matching score distribution. The use of a public dataset permits to apply cohort normalization strategies in contexts in which privacy protection regulations restrict the storage of biometric data. We performed a technological and a scenario evaluation using a commercial matcher currently adopted in real ABC systems and we used data simulating different conditions typical of ABC systems, obtaining encouraging results.

Computational Intelligence (CI) techniques are being increasingly used for automatic monitoring and control systems, especially regarding industrial and environmental applications, due to their performance, their capabilities in fusing noisy or incomplete data obtained from heterogeneous sensors, and the ability in adapting to variations in the operational conditions. Moreover, the increase in the computational power and the decrease of the size of the computing architectures allowed a more pervasive use of CI techniques in a great variety of situations. In this paper, we propose a brief review of the most important CI techniques applied in each step of the design of a monitoring and control system for industrial and environmental applications, and describe how these techniques are integrated in the development of efficient industrial and environmental applications.

Biometric recognition in Automated Border Control (ABC) systems is performed in response to an increased worldwide traffic, by automatically verifying the identity of the passenger during border crossing. Currently, ABC systems seldom use methods for multimodal biometric fusion, which have been proved to increase the recognition accuracy, due to technological and privacy limitations. This paper proposes a framework for the biometric fusion in ABC systems, with the features of being technology-neutral and privacy- compliant, by performing an analysis of the most suitable biometric fusion techniques for ABC systems and considering the current technical and legal limitations.

The analysis of the quality of particulate materials is of great importance for a variety of research and industrial applications. Most image-based methods rely on the segmentation of the image to measure the particles and aggregate their characteristics. However, the segmentation of particulate materials can be severely affected when the setup is not controlled. For instance, when there are device errors, changes in the light conditions, or when the camera gets dirty because of the dust or a similar substance. All of these circumstances are common in industrial setups, like the one studied in this paper. This work presents a framework for quality estimation based on image processing algorithms that avoids segmentation. The considered application scenario is the online quality control of the production of Oriented Strand Boards (OSB), a type of wood panel frequently used in construction and manufacturing industries. The proposed method quantizes frequency domain into a histogram using a non-parametric method, which is later exploited using computational intelligence to classify the quality of superimposed wood particles deposed on a conveyor belt. The method has been tested using synthetic and real images with different noise conditions. The results illustrate the robustness of the approach and its capability to detect significant quality changes in the wood particles.

In this chapter, innovative methods for touchless, less-constrained palmprint recognition are described. In particular, these palmprint recognition systems do not require contact of the hand with any surface or the use of a fixed positioning system. The only constraint imposed on the user is that the hand must be placed open, facing the CCD cameras, and inside the field of view of the cameras. Three-dimensional models are used to compute a representation that is independent of the acquisition position.

Three-dimensional models of fingerprints obtained from contactless acquisitions have the advantages of reducing the distortion present in traditional contact-based samples and the effects of dirt on the finger and the sensor surface. Moreover, they permit to use a greater area for the biometric recognition. The design and test of three-dimensional reconstruction algorithms and contactless recognition methods require the collection of large databases. Since this task can be expensive and timeconsuming, some methods in the literature deal with the generation of synthetic biometric samples. At the best of our knowledge, however, there is only a preliminary study on the computation of small areas of synthetic three-dimensional fingerprints. In this paper, we extend our previous work and describe a virtual environment for the generation of complete threedimensional fingertip shapes, which can be useful for the research community working in the field of three-dimensional fingerprint biometrics. The method is based on image processing techniques and algorithms designed for biometric recognition. We validated the realism of the simulated models by comparing them with real contactless acquisitions. Results show that the method is feasible and produces realistic three-dimensional samples which can effectively be processed by biometric recognition algorithms.

Contactless fingerprint recognition systems are being researched in order to reduce intrinsic limitations of traditional biometric acquisition technologies, encompassing the release of latent fingerprints on the sensor platen, non-linear spatial distortions in the captured samples, and relevant image differences with respect to the moisture level and pressure of the fingertip on the sensor surface.Fingerprint images captured by single cameras, however, can be affected by perspective distortions and deformations due to incorrect alignments of the finger with respect to the camera optical axis. These non-idealities can modify the ridge pattern and reduce the visibility of the fingerprint details, thus decreasing the recognition accuracy. Some systems in the literature overcome this problem by computing three-dimensional models of the finger. Unfortunately, such approaches are usually based on complex and expensive acquisition setups, which limit their portability in consumer devices like mobile phones and tablets. In this paper, we present a novel approach able to recover perspective deformations and improper fingertip alignments in single camera systems. The approach estimates the orientation difference between two contactless fingerprint acquisitions by using neural networks, and permits to register the considered samples by applying the estimated rotation angle to a synthetic three-dimensional model of the finger surface. The generalization capability of neural networks offers a significant advantage by allowing processing a robust estimation of the orientation difference with a very limited need of computational resources with respect to traditional techniques. Experimental results show that the approach is feasible and can effectively enhance the recognition accuracy of single-camera biometric systems. On the evaluated dataset of 800 contactless images, the proposed method permitted to decrease the equal error rate of the used biometric system from 3.04% to 2.20%.

Contactless biometric recognition performed using three-dimensional fingerprint models has the advantages of reducing problems related to the deformations of the skin, dust on the sensor, and spoofing of latent fingerprints. Moreover, the fingerprint area usable for the recognition is wider than the one captured by traditional contact-based acquisition techniques.

Soft biometric techniques can perform a fast and unobtrusive identification within a limited number of users, be used as a preliminary screening filter, or combined in order to increase the recognition accuracy of biometric systems. The weight is a soft biometric trait which offers a good compromise between distinctiveness and permanence, and is frequently used in forensic applications. However, traditional weight measurement techniques are time-consuming and have a low user acceptability. In this paper, we propose a method for a contactless, low-cost, unobtrusive, and unconstrained weight estimation from frame sequences representing a walking person. The method uses image processing techniques to extract a set of features from a pair of frame sequences captured by two cameras. Then, the features are processed using a computational intelligence approach, in order to learn the relations between the extracted characteristics and the weight of the person. We tested the proposed method using frame sequences describing eight different walking directions, and captured in uncontrolled light conditions. The obtained results show that the proposed method is feasible and can achieve a view-independent weight estimation, also without the need of computing a complex model of the body parts.

Traditional biometric systems based on the fingerprint characteristics acquire the biometric samples using touch-based sensors. Some recent researches are focused on the design of touch-less fingerprint recognition systems based on CCD cameras. Most of these systems compute three-dimensional fingertip models and then apply unwrapping techniques in order to obtain images compatible with biometric methods designed for images captured by touch-based sensors. Unwrapped images can present different problems with respect to the traditional fingerprint images. The most important of them is the presence of deformations of the ridge pattern caused by spikes or badly reconstructed regions in the corresponding three-dimensional models. In this paper, we present a neural-based approach for the quality estimation of images obtained from the unwrapping of three-dimensional fingertip models. The paper also presents different sets of features that can be used to evaluate the quality of fingerprint images. Experimental results show that the proposed quality estimation method has an adequate accuracy for the quality classification. The performances of the proposed method are also evaluated in a complete biometric system and compared with the ones obtained by a well-known algorithm in the literature, obtaining satisfactory results.

The estimation of the volume occupied by an object is an important task in the fields of granulometry, quality control, and archaeology. An accurate and well know technique for the volume measurement is based on the Archimedes' principle. However, in many applications it is not possible to use this technique and faster contact-less techniques based on image processing or laser scanning should be adopted.

Iris recognition is nowadays considered as one of the most accurate biometric recognition techniques. However, the overall performances of such systems can be reduced in non-ideal conditions, such as unconstrained, on-the-move, or non-collaborative setups.

In this paper, we propose an image processing system for the detection of wildfire smoke based on computational intelligence techniques and capable of adapting to different applicative environments. The proposed system is designed for processing with limited computational complexity. The detection process focuses on the extraction of specific features of wildfire smoke. A computational intelligence classifier is adopted to identify the presence of smoke. In order to test its effectiveness, the proposed system has been tested with low quality frame sequences, providing the capability to deal also with low cost cameras. The results indicate that the proposed approach is accurate and can be effectively applied in different environmental conditions.

Current contactless fingertip recognition systems based on three-dimensional finger models mostly use multiple views (N > 2) or structured light illumination with multiple patterns projected over a period of time. In this paper, we present a novel methodology able to obtain a fast and accurate three-dimensional reconstruction of the fingertip by using a single two-view acquisition and a static projected pattern. The acquisition setup is less constrained than the ones proposed in the literature and requires only that the finger is placed according to the depth of focus of the cameras, and in the overlapping field of views. The obtained pairs of images are processed in order to extract the information related to the fingertip and the projected pattern. The projected pattern permits to extract a set of reference points in the two images, which are then matched by using a correlation approach. The information related to a previous calibration of the cameras is then used in order to estimate the finger model, and one input image is wrapped on the resulting three-dimensional model, obtaining a three-dimensional pattern with a limited distortion of the ridges. In order to obtain data that can be treated by traditional algorithms, the obtained three-dimensional models are then unwrapped into bidimensional images. The quality of the unwrapped images is evaluated by using a software designed for contact-based fingerprint images. The obtained results show that the methodology is feasible and a realistic three-dimensional reconstruction can be achieved with few constraints. These results also show that the fingertip models computed by using our approach can be processed by both specific three-dimensional matching algorithms and traditional matching approaches. We also compared the results with the ones obtained without using structured light techniques, showing that the use of a projector achieves a faster and more accurate fingertip reconstruction.

Biometric systems identify individuals by comparison of the individual biometric traits, such as the fingerprint patterns. In the literature, many relevant methods are based on the localization of a reference “pivot” point of the fingerprint, called principal singular point (PSP). Most of the time, the PSP is selected from the list of the estimated singular points (SPs) that are identified by specific local patterns of the fingerprint ridges, called cores and deltas. The challenge is to provide an automatic method capable to select the same PSP from different images of the same fingertip. In this paper, we propose a technique that estimates the position of all the singular points by processing the global structure of the ridges and extracting a specific set of features. The selection of the reference point from the candidate list is then obtained by processing the extracted features with computational intelligence classification techniques. Experiments show that the method is accurate and it can be applied on contact and contact-less image types.

For more than a century, the methods for data representation and the exploration of the intrinsic structures of data have developed remarkably and consist of supervised and unsupervised methods. However, recent years have witnessed the flourishing of big data, where typical dataset dimensions are high and the data can come in messy, incomplete, unlabeled, or corrupted forms. Consequently, discovering the hidden structure buried inside such data becomes highly challenging. From this perspective, exploratory data analysis plays a substantial role in learning the hidden structures that encompass the significant features of the data in an ordered manner by extracting patterns and testing hypotheses to identify anomalies. Unsupervised generative learning models are a class of machine learning models characterized by their potential to reduce the dimensionality, discover the exploratory factors, and learn representations without any predefined labels; moreover, such models can generate the data from the reduced factors’ domain. The beginner researchers can find in this survey the recent unsupervised generative learning models for the purpose of data exploration and learning representations; specifically, this article covers three families of methods based on their usage in the era of big data: blind source separation, manifold learning, and neural networks, from shallow to deep architectures.