Search

Visible watermarking offers direct copyright protection to the naked eye using visible pattern superimposed in the host image. However, due to their exposure, the watermark pattern becomes an easy target to remove by malicious entities that generates copyright-free images. Until now, in the scientific literature several semi-automatic visible watermark removal attacks have been proposed, showing enough visual quality of the recovered host image. In this paper, we propose a watermark design based on modified Seam Carving technique to provide robustness against removal attacks including inpainting-based attacks, Independent Component Analysis (ICA)-based attacks, and image matting-based attacks. Robustness against the attacks is measured using common image quality assessments and a watermark visibility assessment based on the Human Visual System (HVS). Experimental results show robustness of the proposed method against above mentioned watermark removal attacks under the most severe situation, in which adversary has the logotype used as a watermark pattern because it is public in Internet. The proposed method can be applied to any visible watermarking algorithms to make them robust against illegal removal attempts, without sacrificing their original functionality, such as reversibility for authorized users.

Cryptographic algorithms have been widely used in all secure data transmissions due their capabilities to hide sensible information against eavesdroppers. Nevertheless, technological progress tends to develop new cryptographic methods based on non-classical computing paradigms. In this work, the first circuit of visual cryptography based on Spiking Neural P systems (VCSN P) is presented. The proposed circuit makes use of new biological behaviors in astrocytes, which are involved in processes of plasticity and pruning control in dendrites, that can be the beginning to new cryptographic algorithms based on SN P systems. To validate the SN P circuit, this was implemented on the Cyclone V field programmable gate array (FPGA) development kit. The results demonstrate that this new approach exploits the features of the SN P systems to improve the cryptographic algorithm.

Human health is threatened by several diseases for this reason automated medical diagnosis systems has been developed several years ago. These systems need databases, the creation of these databases is tedious, arduous and stops being done so the created database is incomplete or unbalanced. Sometimes the databases are private to protect the private information of the patients, among other problems. For this reason, the researchers have started to use synthetic data. The synthetic data have been applied by different hospitals in the USA. The creation of synthetic data has different problems like the synthetic data are generated using rules defined by the user, the proposed approaches only can create one kind of data, the proposals require input from domain experts, among others. To address these kinds of problems, we propose a novel approach, which consists of the Bidirectional Recurrent Neural Network and the statistical stage to generate synthetic biomedical signals. The approach is able to create 5 kinds of biomedical signals (ECG, EEG, BCG, PPG, and Respiratory Impedance). Our approach is able to create synthetic data for patients or for specific events. The performance of our approach is compared with other generative models (GAN’s) through evaluation metrics. The created synthetic data are used to construct models; these models are able to successfully differentiate between different signals with high accuracies.

Nowadays, the administration of medical images can be compromised by several information security issues. Among others, a first problem consists in authenticating the source that originated the medical image and verifying that it corresponds to the correct patient; a second problem is aimed to avoid detachment between an electronic patient record and its corresponding images. A promising solution is the use of digital watermarking techniques applied to medical images considering in the design requirements of imperceptibility-robustness of the field of medical imaging. In this paper a digital watermarking method for an efficient administration of ultrasound images is proposed. To avoid the detachment between the record and images, we use an invisible watermarking in the frequency domain of the image. To authenticate the source of origin of the medical image, we use a visible-imperceptible watermarking in the spatial domain of the image. The experimental results show an efficient robustness and imperceptibility of the embedded watermark signals into the medical image. A comparison of performance with related works previously reported in the literature is included.

Heart disease is the principal cause of death across the globe and the ECG signals are used to diagnose it. The correct classification of this disease allows us the opportunity to apply a more focused treatment. ECG signals are fed into Automated Diagnosis Systems, and these systems use techniques like processing digital signals, machine learning, and deep learning. This paper shows the results when the sampling frequency of the ECG signals is resampled and proposes a new preprocessing stage. The new stage applies Wavelet based on Atomic Functions to eliminate the noise and baseline wander. The Wavelet based on Atomic Functions have demonstrated successful performances in computer science. The ECG signals are segmented into 1, 2, 5, and 10 s; these segmented signals are fed into the classifier stage. Our proposal was tested in four accessible public databases separately, and finally by gathering the databases. We were able to successfully differentiate between 11 types of ECG signals with an accuracy of 98.9%.

One of the most significant current discussions in Computer Science is the skin recognition. Many papers have studied the skin detection using different techniques, artificial vision, machine learning, deep learning among others. Despite its long success, the skin recognition has several problems in use. However, there has been little discussion about generate a skin tonalities classification. The aim of this paper is to propose a system to skin tonalities through geographic distribution based on clustering algorithms, pattern recognition and fuzzy logic. This distribution gives us the opportunity to classify the skin tonalities. We can study each skin tonality for any applications as medical diagnosis, security. In the first stage, we use the RGB color model to training the system. Then, we tested the system with different color models. We use color model with the best result to propose geographic distribution based skin tonalities. The results show that is possible to generate a skin tonalities classification. The proposed system is using to skin recognition, showing interesting results under controlled and no controlled conditions.

This paper proposes the use of the FASSD-Net model for semantic segmentation of human silhouettes, these silhouettes can later be used in various applications that require specific characteristics of human interaction observed in video sequences for the understanding of human activities or for human identification. These applications are classified as high-level task semantic understanding. Since semantic segmentation is presented as one solution for human silhouette extraction, it is concluded that convolutional neural networks (CNN) have a clear advantage over traditional methods for computer vision, based on their ability to learn the representations of appropriate characteristics for the task of segmentation. In this work, the FASSD-Net model is used as a novel proposal that promises real-time segmentation in high-resolution images exceeding 20 FPS. To evaluate the proposed scheme, we use the Cityscapes database, which consists of sundry scenarios that represent human interaction with its environment (these scenarios show the semantic segmentation of people, difficult to solve, that favors the evaluation of our proposal), To adapt the FASSD-Net model to human silhouette semantic segmentation, the indexes of the 19 classes traditionally proposed for Cityscapes were modified, leaving only two labels: One for the class of interest labeled as person and one for the background. The Cityscapes database includes the category “human” composed for “rider” and “person” classes, in which the rider class contains incomplete human silhouettes due to self-occlusions for the activity or transport used. For this reason, we only train the model using the person class rather than human category. The implementation of the FASSD-Net model with only two classes shows promising results in both a qualitative and quantitative manner for the segmentation of human silhouettes.

At present, new data sharing technologies, such as those used in the Internet of Things (IoT) paradigm, are being extensively adopted. For this reason, intelligent security controls have become imperative. According to good practices and security information standards, particularly those regarding security in depth, several defensive layers are required to protect information assets. Within the context of IoT cyber-attacks, it is fundamental to continuously adapt new detection mechanisms for growing IoT threats, specifically for those becoming more sophisticated within mesh networks, such as identity theft and cloning. Therefore, current applications, such as Intrusion Detection Systems (IDS), Intrusion Prevention Systems (IPS), and Security Information and Event Management Systems (SIEM), are becoming inadequate for accurately handling novel security incidents, due to their signature-based detection procedures using the matching and flagging of anomalous patterns. This project focuses on a seldom-investigated identity attack—the Clone ID attack—directed at the Routing Protocol for Low Power and Lossy Networks (RPL), the underlying technology for most IoT devices. Hence, a robust Artificial Intelligence-based protection framework is proposed, in order to tackle major identity impersonation attacks, which classical applications are prone to misidentifying. On this basis, unsupervised pre-training techniques are employed to select key characteristics from RPL network samples. Then, a Dense Neural Network (DNN) is trained to maximize deep feature engineering, with the aim of improving classification results to protect against malicious counterfeiting attempts.

This paper proposes a new symmetric encryption system based on an elliptical curve and chaos, where the encryption is done in a single block and runs for 14 rounds. Here, the 15 encryption keys have the same size as the image and are generated using a solution point of a strong elliptic curve. Using a string of random numbers obtained with a logistic map, a permutation and its inverse are generated, which improve the encryption level and add diffusion to the cryptosystem. Another important contribution to this research is the generation of a substitution box with a non-linearity of 100, which strengthens the cryptosystem against differential and linear attacks that add confusion to the scheme. Moreover, the cryptographic properties of the proposed S-Box are compared with those of the S-Box of the Advanced Encryption Standard (AES) to ascertain that it is a suitable alternative that it is resistant to differential power analysis (DPA) attacks. To verify the robustness of proposed cryptosystem against cryptanalysis and the quality of the cipher text obtained, this system is subjected to different tests, such as entropy, correlation coefficient, χ2, Changing Pixel Rate (NPCR), and Unified Averaged Changing Intensity (UACI). The results are shown and compared with recently proposed systems.

COVID-2019 is a global threat, for this reason around the world, researches have been focused on topics such as to detect it, prevent it, cure it, and predict it. Different analyses propose models to predict the evolution of this epidemic. These analyses propose models for specific geographical areas, specific countries, or create a global model. The models give us the possibility to predict the virus behavior, it could be used to make future response plans. This work presents an analysis of COVID-19 spread that shows a different angle for the whole world, through 6 geographic regions (continents). We propose to create a relationship between the countries, which are in the same geographical area to predict the advance of the virus. The countries in the same geographic region have variables with similar values (quantifiable and non-quantifiable), which affect the spread of the virus. We propose an algorithm to performed and evaluated the ARIMA model for 145 countries, which are distributed into 6 regions. Then, we construct a model for these regions using the ARIMA parameters, the population per 1M people, the number of cases, and polynomial functions. The proposal is able to predict the COVID-19 cases with a RMSE average of 144.81. The main outcome of this paper is showing a relation between COVID-19 behavior and population in a region, these results show us the opportunity to create more models to predict the COVID-19 behavior using variables as humidity, climate, culture, among others.

This paper proposes a teleophthalmology support system in which we use algorithms of object detection and semantic segmentation, such as faster region-based CNN (FR-CNN) and SegNet, based on several CNN architectures such as: Vgg16, MobileNet, AlexNet, etc. These are used to segment and analyze the principal anatomical elements, such as optic disc (OD), region of interest (ROI) composed by the macular region, real retinal region, and vessels. Unlike the conventional retinal image quality assessment system, the proposed system provides some possible reasons about the low-quality image to support the operator of an ophthalmoscope and patient to acquire and transmit a better-quality image to central eye hospital for its diagnosis. The proposed system consists of four steps: OD detection, OD quality analysis, obstruction detection of the region of interest (ROI), and vessel segmentation. For the OD detection, artefacts and vessel segmentation, the FR-CNN and SegNet are used, while for the OD quality analysis, we use transfer learning. The proposed system provides accuracies of 0.93 for the OD detection, 0.86 for OD image quality, 1.0 for artefact detection, and 0.98 for vessel segmentation. As the global performance metric, the kappa-based agreement score between ophthalmologist and the proposed system is calculated, which is higher than the score between ophthalmologist and general practitioner.

This work addresses the use of the MOTIF algorithm for face feature extraction. The MOTIF algorithm is commonly used to characterize texture and shows good performance in this task; a MOTIF algorithm without the Co-occurrence Matrix is proposed to obtain face features, and the approach proves to be effective. System testing was based on a standard database (the AR Face database) that includes 120 people, 70 images with face expressions and 30 with sunglasses; 1 to 9 images were used to make the template for each person. After using Euclidean distance, Cosine distance and support vector machine as classifiers, correct classification was achieved with 98% accuracy. Further tests were performed with all databases and compared with Local Binary Pattern, DI-WBP and other commonly used schemes, demonstrating effective face recognition by the MOTIF algorithm without the co-occurrence matrix in addition to its fast performance due to the low computational cost.

In recent years, end users can easily capture digital images using several devices, such as smartphones, mobile devices and digital imaging cameras, allowing such images to be easily copied, manipulated, transmitted or format converted without any restrictions. This fact suggests the necessity to develop digital tools, such as digital watermarking, to solve the issues associated with copyright protection and ownership authentication of digital images. To claim the ownership of a digital image, we propose a camouflaged, unseen-visible watermarking technique based on luminance and texture properties in conjunction with an image enhancement criterion. The proposed method has some advantages over invisible and visible watermarking methodologies in terms of readability and imperceptibility of the watermark, respectively. The experimental results demonstrate that the proposed scheme is effective and applicable for digital images on a variety of topics, including natural scenes and man-made objects, both indoors and outdoors. A comparison with previously reported methods based on unseen-visible watermarking techniques is also provided.

An imperceptible visible watermarking (IVW) algorithm is proposed to deliver auxiliary information about the visual contents, including ownership information. The proposed algorithm overcomes several inconveniences presented in previously reported IVW algorithms using the just noticeable difference (JND) criterion, invisible watermarking and binarisation function. The proposed algorithm consists of the visible watermark embedding and the exhibition stages. The JND criterion is used to embed visible watermark pattern in an imperceptible manner by the human visual system, while invisible watermarking based on the discrete cosine transform is used to share crucial parameters between both stages, which makes possible the watermark visualisation without any side information. In the watermark exhibition stage, the imperceptibly embedded visible watermark pattern is exhibited by the binarisation function. The proposed algorithm can be applied to any class of images with different characteristics and several visible watermark patterns can be embedded into an input image. Evaluation results show that performance of the proposed algorithm is better compared with previously reported algorithms from several practical points of view.

During the last two decades, several visible watermarking algorithms with different characteristics have been proposed. In the context of visible digital watermarking, watermark visibility and its non-obtrusiveness are essential issues. However, assessment metrics were not established to evaluate these two issues until now. Due to the lack of established assessment metrics, visibility is evaluated in a subjective manner and the obtrusiveness caused by the visible watermark is evaluated by well-known image quality metrics, such as peak signal noise ratio, comparing the host image and the watermarked image. In this paper, we propose four assessment metrics for visible watermarking algorithms, in which the global visibility, global obtrusiveness, local obtrusiveness in host edge region, and a global quality of watermarked image are evaluated using a couple of criteria based on the human visual system and a pixel-based just-noticeable distortion function. The evaluation results show the reliability of the proposed metrics used for measuring the above-mentioned aspects when several visible watermarking algorithms are employed as well as in several scenarios where an attempt to remove the watermark pattern from the watermarked content was made.

Some watermarking authentication methods based on spectral domain using QR codes and Quantization Index Modulation (QIM) are proposed for ownership protection in color images. The QR code is created with the owners’ information and used as binary watermark sequence, which is permuted using Arnold Permutation to increase the security before embedding it. Once the watermark sequence is generated, the original color image is transformed from RGB to YCbCr color space where the Luminance Channel (Y) is processed by the Singular Value Decomposition (SVD), Discrete Wavelet Transform (DWT) and Discrete Cosine Transform (DCT) to embed the binary watermark sequence using Quantization Index Modulation (QIM). The experimental results show the effectiveness of the proposed method.

This paper proposes a removable visible watermarking system based on a dual watermark technique and blind removal. A visible watermark pattern is embedded in the cosine discrete transform (DCT) domain, taking into consideration the texture and luminance features of the watermark and host images to create a visible watermarked image. To prevent illegal visible watermark removal, the original watermark is embedded in an invisible manner in the visible watermarked image by employing the Quantization Index Modulation-Dither Modulation (QIM-DM) technique, thus ensuring that the original watermark cannot be obtained by malicious attacks. The visible watermark removal process is carried out using only the correct user’s keys, without the need for additional information, such as the original watermark or the original host image, which allows a high-quality image to be obtained; however, if the user’s keys used in the removal process are wrong, the visible watermarked image suffers higher distortion in its content, even in non-visible watermarked regions. The experimental results show that the proposed system outperforms previous related works in terms of blind removal, preservation of the quality of the unmarked recovered image, and higher visual degradation of the content in the recovered image if an illegal removal attempt is performed.

Presently, security is a hot research topic due to the impact in daily information infrastructure. Machine-learning solutions have been improving classical detection practices, but detection tasks employ irregular amounts of data since the number of instances that represent one or several malicious samples can significantly vary. In highly unbalanced data, classification models regularly have high precision with respect to the majority class, while minority classes are considered noise due to the lack of information that they provide. Well-known datasets used for malware-based analyses like botnet attacks and Intrusion Detection Systems (IDS) mainly comprise logs, records, or network-traffic captures that do not provide an ideal source of evidence as a result of obtaining raw data. As an example, the numbers of abnormal and constant connections generated by either botnets or intruders within a network are considerably smaller than those from benign applications. In most cases, inadequate dataset design may lead to the downgrade of a learning algorithm, resulting in overfitting and poor classification rates. To address these problems, we propose a resampling method, the Synthetic Minority Oversampling Technique (SMOTE) with a grid-search algorithm optimization procedure. This work demonstrates classification-result improvements for botnet and IDS datasets by merging synthetically generated balanced data and tuning different supervised-learning algorithms.

The counting of vehicles plays an important role in measuring the behavior patterns of traffic flow in cities, as streets and avenues can get crowded easily. To address this problem, some Intelligent Transport Systems (ITSs) have been implemented in order to count vehicles with already established video surveillance infrastructure. With this in mind, in this paper, we present an on-line learning methodology for counting vehicles in video sequences based on Incremental Principal Component Analysis (Incremental PCA). This incremental learning method allows us to identify the maximum variability (i.e., motion detection) between a previous block of frames and the actual one by using only the first projected eigenvector. Once the projected image is obtained, we apply dynamic thresholding to perform image binarization. Then, a series of post-processing steps are applied to enhance the binary image containing the objects in motion. Finally, we count the number of vehicles by implementing a virtual detection line in each of the road lanes. These lines determine the instants where the vehicles pass completely through them. Results show that our proposed methodology is able to count vehicles with 96.6% accuracy at 26 frames per second on average&mdash, dealing with both camera jitter and sudden illumination changes caused by the environment and the camera auto exposure.

Due to the vertiginous growth of malicious actors, malware has been crafted, distributed and propagated around the world with new and sophisticated techniques. Classical malware detection procedures, mostly based on signatures and heuristic searches, are now being replaced with machine learning-based (ML) solutions. However, some challenges are still present. Firstly, supervised approaches use anti-virus tags to create hand-crafted datasets, resulting in a lack of taxonomy and uncertainty if a given observation is classified with a proper label. Secondly, off-line and feed-forward approaches may result in complex and time consuming feature extraction tasks. In this work, we propose a novel method that reinforces malware characterization by capturing rich relevance and contextual patterns into an n-dimensional weighted word embedding vector (WEV) space. Results prove that by clustering similar WEVs via unsupervised learning, malware can be categorized into four major families, improving detection with less resources.

In recent years, Online Social Networks (OSNs) have received a great deal of attention for their potential use in the spatial and temporal modeling of events owing to the information that can be extracted from these platforms. Within this context, one of the most latent applications is the monitoring of natural disasters. Vital information posted by OSN users can contribute to relief efforts during and after a catastrophe. Although it is possible to retrieve data from OSNs using embedded geographic information provided by GPS systems, this feature is disabled by default in most cases. An alternative solution is to geoparse specific locations using language models based on Named Entity Recognition (NER) techniques. In this work, a sensor that uses Twitter is proposed to monitor natural disasters. The approach is intended to sense data by detecting toponyms (named places written within the text) in tweets with event-related information, e.g., a collapsed building on a specific avenue or the location at which a person was last seen. The proposed approach is carried out by transforming tokenized tweets into word embeddings: a rich linguistic and contextual vector representation of textual corpora. Pre-labeled word embeddings are employed to train a Recurrent Neural Network variant, known as a Bidirectional Long Short-Term Memory (biLSTM) network, that is capable of dealing with sequential data by analyzing information in both directions of a word (past and future entries). Moreover, a Conditional Random Field (CRF) output layer, which aims to maximize the transition from one NER tag to another, is used to increase the classification accuracy. The resulting labeled words are joined to coherently form a toponym, which is geocoded and scored by a Kernel Density Estimation function. At the end of the process, the scored data are presented graphically to depict areas in which the majority of tweets reporting topics related to a natural disaster are concentrated. A case study on Mexico&rsquo, s 2017 Earthquake is presented, and the data extracted during and after the event are reported.

Area consumption is one of the most important design constrains in the development of compact digital systems. Several authors have proposed making compact Cochlear Implant processors using Gamma-tone filter banks. These model aspects of the cochlea spectral filtering. A good area-efficient design of the Gamma-tone Filter Bank could reduce the amount of circuitry allowing patients to wear these cochlear implants more easily. In consequence, many authors have reduced the area by using the minimum number of registers when implementing this type of filter. However, critical paths limit their performance. Here a compact Gamma-tone Filter processor, formulated using the impulse invariant transformation together with a normalization method, is presented. The normalization method in the model guarantees the same precision for any filter order. In addition, area resources are kept low due to the implementation of a single Second Order Section (SOS) IIR stage for processing several SOS IIR stages and several channels at different times. Results show that the combination of the properties of the model and the implementation techniques generate a processor with high processing speed, expending less resources than reported in the literature.

In this article, an adaptive beamforming system based on the augmented complex least mean square algorithm is analysed. In this approach, the adaptive filter is used as a widely linear system. The second-order statistical information of the signals involved in the array is exploited. Under this consideration, the ability of the adaptive array to minimize the effects of interferences and complex white noise could be enhanced. The equations for the optimal weights and the array factor are derived for the proposed beamforming system. Computer simulations have been performed to evaluate the performance of the adaptive array, and the results were compared with two of the most common standard adaptive beamforming algorithms: the least mean square and recursive least square. The numerical simulations show that the proposed adaptive array has a better performance in time and spatial domain as compared to the classical beamforming systems.

The development of coding schemes with the capacity to simultaneously encrypt and compress audio signals is a subject of active research because of the increasing necessity for transmitting sensitive audio information over insecure communication channels. Thus, several schemes have been developed, firstly, some of them compress the digital information and subsequently encrypt the resulting information. These schemas efficiently compress and encrypt the information. However, they may compromise the information as it can be accessed before encryption. To overcome this problem, a compressing sensing-based system to simultaneously compress and encrypt audio signals is proposed in which the audio signal is segmented in frames of 1024 samples and transformed into a sparse frame using the discrete cosine transform (DCT). Each frame is then multiplied by a different sensing matrix generated using the chaotic mixing scheme. This fact allows that the proposed scheme satisfies the extended Wyner secrecy (EWS) criterion. The evaluation results obtained using several genres of audio signals show that the proposed system allows to simultaneously compress and encrypt audio signals, satisfying the EWS criterion.

Malware analysis using machine learning techniques has been the subject of study in recent years as a new alternative for efficient detection of malicious behaviour patterns in different operating systems. Recent advances in this research field have proposed different algorithms employing information extraction and feature selection tasks, aiming to cover different types of data and improving several performance metrics. In this work is proposed the use of an assembly classifier, better known as Random Forest, that improves the performance of other well-known algorithms by aggregating individual class predictions to combine into a final prediction. A case study is presented using two different datasets of malware, that through data preparation techniques is enhanced the quality of data to strengthen the classifier training.

In this paper we propose a security enhancement method for color images management. Our proposed algorithm is a combination of invisible watermarking and reversible data-hiding, both performed in the encrypted domain of the color images. Initially, the original image is converted from RGB to YCbCr color model and then each component is ciphered by applying a chaotic mixing procedure by using a first content owner's key. Reversible data hiding technique is used to embed additional bits into the ciphered chrominance information. With purposes of ownership authentication, an invisible watermarking technique is used to embed a watermark into the ciphered luminance information. Finally, to increase the security of the method, the ciphered chrominance and luminance components with hidden data are ciphered again by using a second content owner's key and a lightweight encryption algorithm. Three options are available for receivers who hold different keys: a) extract embedded additional bits and the watermark, b) decrypt the watermarked image with embedded additional bits, or c) extract the additional bits, the watermark and generate the recovered image that is identical to the original, respectively. Experimental results show the imperceptibility, capacity and security of the proposed method in the context of watermarking and reversible data hiding techniques, respectively.

Secret sharing (SS) schemes based on cellular automata (CA) are considered as secure encrypting algorithms, where several secret data can be shared among some persons. Recently the SS schemes can be applied to solve real-world problems, such as security in cloud computing. The principal obstacle of use of the SS scheme is its considerably high computational cost; especially if a large amount of secret data must be encrypted and shared. In this work, we propose a parallel CA-based SS scheme suitable for any kinds of digital data in the graphic processing unit using compute unified device architecture technology. The uses of global memory and shared memory are analyzed from computational effectiveness and security points of view. The experimental results show the proposed parallel implementation provides a speedup rate more than 18-fold compared with its sequential implementation. Also we show the increase of the security level of the parallel implementation with respect to the sequential implementation.

Secret image sharing (SIS) techniques allow visual secrets to be shared between multiple people. These techniques require a predetermined access structure to be satisfied to reveal the secret. SIS schemes have the potential to increase security in several applications including telemedicine and image transfer in the cloud computing environment, providing controlled access to confidential images. To date, a significant number of SIS schemes with various properties and access structures have been proposed. Among them, hierarchical SIS (HSIS) is considered foremost since almost all organizations and associations manage their secret information in a hierarchical manner. However, the HSIS scheme tends to suffer from security flaws as the secret image can often be visually leaked, even when the access structure is not properly satisfied. To avoid this security flaw, we propose a cheating prevention mechanism by introducing a robust watermarking technique based on the Quantization Index Modulation-Dither Modulation (QIM-DM) algorithm in the discrete cosine transform (DCT) domain. Experimental results demonstrate the superior effectiveness of our proposed cheating prevention mechanism. Furthermore, our approach retains all of the desirable properties of the HSIS scheme.

This paper proposes a view-invariant gait recognition algorithm, which builds a unique view invariant model taking advantage of the dimensionality reduction provided by the Direct Linear Discriminant Analysis (DLDA). Proposed scheme is able to reduce the under-sampling problem (USP) that appears usually when the number of training samples is much smaller than the dimension of the feature space. Proposed approach uses the Gait Energy Images (GEIs) and DLDA to create a view invariant model that is able to determine with high accuracy the identity of the person under analysis independently of incoming angles. Evaluation results show that the proposed scheme provides a recognition performance quite independent of the view angles and higher accuracy compared with other previously proposed gait recognition methods, in terms of computational complexity and recognition accuracy.

This paper presents a new method for movement detection, based on the projection of the first eigenvalue obtained by using the Incremental Principal Component Analysis (IPCA). This is obtained assuming that the incremental computation of the eigenvalues using two images, simultaneously, produces an eigenvector that projects the maximum variability among both images and then, the movement can be estimated.

In this paper, we consider the problem of enhancing the filtering capacity of single standard beamformers for real base-band signals. Unfortunately, the standard smart arrays are not able to effectively minimize the effects of the complex noise in the array. In order to solve this problem, we analyze a parallel adaptive beamforming system based on complex nonlinear adaptive filters. In this design, the smart array is used as a widely linear (WL) system in which the second order statistical information of the signals is exploited by using the covariance and pseudo-covariance matrices. The simulations results show this architecture has a lead filtering capability as well as an outstanding generation of the radiation patterns. Furthermore, this parallel hybrid array showed to have a better filtering capacity than the standard arrays LMS, NLMS and RLS. General equations for the optimal weight vector are derived for this augmented complex (AC) smart array or widely linear beamforming.

Dengue, Chikungunya and Zika viruses cause dangerous infections in tropical and subtropical regions throughout the world. The World Health Organization estimates that one out of every three persons in the entire human population is in danger of contracting one of these diseases from a single mosquito bite. Currently, these viral infections are not preventable by vaccines and there is not a direct treatment that can effectively diminish the viral infection, which causes a wide range of pathologies, including severe joint pain, internal blood loss, permanent neurological damage in unborn children and even death. Due to this grim scenario, the best and maybe the only line of defense against these diseases is the effective surveillance, control and suppression of the mosquitoes that transmit these viruses: Aedes aegypti and Aedes albopictus. In this paper, we present a complete solution that is capable of identifying the Aedes aegypti and Aedes albopictus mosquito in the larval stage, which is easily disposable, restricted to water bodies, and incapable of transmitting diseases according to the Centers for Disease Control and Prevention (CDC). Our proposal is based on deep neural networks (DNN) that effectively recognize larval samples with an accuracy of 94.19%, which is better than other state-of-the-art automatic methods. Additionally, the capabilities of our proposed DNN allow us to automatically crop the region of interest (ROI) with an accuracy of 92.85% and then automatically classify the region as Aedes positive or Aedes negative, without further human intervention and in less than a second, accelerating the response time for biological control from days to seconds. Our proposal includes hardware designs that allow inexpensive implementation, making it suitable for isolated communities, underdeveloped countries, and rural or urban areas.

Nowadays, the management of medical imaging could be affected by several issues related to information security. Among others, a critical issue is related to information authentication. A promissory way to solve this issue is the use of digital watermarking to improve the management of medical imaging. In this paper, we propose a robust and secure medical image watermarking method that involves invisible and zero watermarking techniques that provide a proper link between clinical information, medical image, and patient identity. Concealing an encoded and encrypted watermark signal in the frequency domain of the medical image, the invisible watermarking component avoids the detachment between the medical images and their corresponding electronic patient record. Based on a modified spread spectrum technique, the response of the invisible watermarking detector and the watermark information, the zero watermarking stage authenticates in a confidential form the identity of the patient. Experimental results show the effectiveness of the proposed method in terms of transparency, security, reliability, and robustness of the watermark against several image processing operations. A comparison with state of the art proposals with similar purposes to the proposed method is also provided. The presented work may find applications in the secure and effective management of medical imaging.

This paper proposes a methodology for early fire detection based on visual smoke characteristics such as movement, color, gray tones and dynamic texture, i.e., diverse but representative and discriminant characteristics, as well as its ascending expansion, which is sequentially processed to find the candidate smoke regions. Thus, once a region with movement is detected, the pixels inside it that are smoke color are estimated to obtain a more detailed description of the smoke candidate region. Next, to increase the system efficiency and reduce false alarms, each region is characterized using the local binary pattern, which analyzes its texture and classifies it by means of a multi-layer perceptron. Finally, the ascending expansion of the candidate region is analyzed and those smoke regions that maintain or increase their ascending growth over a time span are considered as a smoke regions, and an alarm is triggered. Evaluations were performed using two different classifiers, namely multi-Layer perceptron and the support vector machine, with a standard database smoke video. Evaluation results show that the proposed system provides fire detection accuracy of between 97.85% and 99.83%.

This paper proposes a view-invariant gait recognition framework that employs a unique view invariant model that profits from the dimensionality reduction provided by Direct Linear Discriminant Analysis (DLDA). The framework, which employs gait energy images (GEIs), creates a single joint model that accurately classifies GEIs captured at different angles. Moreover, the proposed framework also helps to reduce the under-sampling problem (USP) that usually appears when the number of training samples is much smaller than the dimension of the feature space. Evaluation experiments compare the proposed framework’s computational complexity and recognition accuracy against those of other view-invariant methods. Results show improvements in both computational complexity and recognition accuracy.

This paper proposes an image-adaptive steganalysis algorithm for LSB-Matching steganography. In the proposed method, the content features of the input image are analyzed in order to obtain a classification composed by plain and non-plain regions. In this context, only the plain regions of the input image are considered to determine if the image contains or not hidden information. The experimental results show that the proposed steganalysis algorithm outperforms the conventional LSB-Matching steganalysis algorithms in terms of detection performance and the same time minimizes the inter-database error. The results are shown through the ROC curves that illustrate a better performance of the proposed algorithm.

This paper proposes a facial expression recognition algorithm that automatically detects the facial image contained in a color picture and segments it in two regions of interest (ROI)—the forehead/eyes and the mouth—which are then divided into non-overlapping N × M blocks. Next, the average of the first element of the cross correlation between 54 Gabor functions and each one of the N × M blocks is estimated to generate a matrix of dimension L × NM , where L is the number of training images. This matrix is then inserted into a principal component analysis (PCA) module for dimensionality reduction. Finally, the resulting matrix is used to generate the feature vectors, which are inserted into the proposed low complexity classifier based on clustering and fuzzy logic techniques. This classifier provides recognition rates close to those provided by other high performance classifiers, but with far less computational complexity. The experimental results show that proposed system achieves a recognition rate of about 97% when the feature vector from only one ROI is used, and that the recognition rate increases to approximately 99% when the feature vectors of both ROIs are used. This result means that the proposed method can achieve an overall recognition rate of approximately 97% even when one of the two ROIs is totally occluded.

In this paper we propose a watermarking scheme to ensure copyright protection in video distribution systems. This research is focused on provide a low computational cost solution that allows embedding a robust and imperceptible watermarking signal in the video content, such that the original video transmission rate will not be affected. Our proposal accomplishes this challenge by introducing a fast and effective method based on spatio-temporal Human Visual System properties that allow improving imperceptibility along video sequence and at the same time, enhancing robustness. Extensive experiments were performed to prove that the proposed algorithm satisfies quickness, invisibility and robustness against several real-world video processing issues. The proposed scheme has the advantages of robustness, simplicity and flexibility, allowing it to be considered a good solution in practical situations since it can be adapted to different video distribution systems.

Este trabajo presenta dos modificaciones del metodo de eigenphases para aumentar su precision. En la primera modificacion llamada Local Spatial Domain Eigenphase (LSDE), la imagen del rostro se divide en bloques de N × N pixeles, cuyas magnitudes se normalizan. Estos bloques se concatenan antes de que el espectro de fase y el PCA se estimen. En la segunda modificacion llamada Local Frecuency Domain Eigenphase (LFDE), despues de la segmentacion de la imagen en bloques de N x N pixeles, las magnitudes de los pixeles de dichos bloques se normalizan y se calcula el espectro de fase en forma independiente. Una vez que se obtiene el espectro de fase de todos los bloques, se concatenan y se procede a la aplicacion del analisis de componentes principales (PCA) para reducir la dimensionalidad del problema. Las modificaciones propuestas se evaluan en la modalidad de identificacion, tanto en "open set" como en "closed-set", asi tambien en verificacion de identidad. En ambos casos se empleo la base de datos AR Face Database. Los resultados experimentales muestran que las modificaciones propuestas presentan un funcionamiento adecuado bajo diferentes condiciones de iluminacion y oclusion parcial

In this paper, we propose a new view-invariant framework for gait analysis. The framework profits from the dimensionality reduction advantages of Direct Linear Discriminant Analysis (DLDA) to build a unique view-invariant model. Among these advantages is the capability to tackle the under-sampling problem (USP), which commonly occurs when the number of dimensions of the feature space is much larger than the number of training samples. Our framework employs Gait Energy Images (GEIs) as features to create a single joint model suitable for classification of various angles with high accuracy. Performance evaluations shows the advantages of our framework, in terms of computational time and recognition accuracy, as compared to state-of-the-art view-invariant methods.