Raffaele Pizzolante

Cloud-based adaptive compression and secure management services for 3D healthcare data

Several studies show that the lack of access to resources and shared data is one of the main causes of errors in the healthcare sector. In particular, 3D medical images play a fundamental role in healthcare environment, but they are typically very large in size. Therefore, their management, which should be performed also by means of devices with limited characteristics, requires complex network protocols along with advanced compression and security techniques. This work concerns the secure management of 3D medical images, with the main aim that such management must take place in an almost completely transparent manner for the end-user, regardless of the computational and networking capabilities he may use. In particular, our contribution is twofold: first, we propose an engine for lossless dynamic and adaptive compression of 3D medical images, which also allows the embedding of security watermarks within them. Furthermore, in order to provide effective, secure and flexible access to healthcare resources that need to be managed by medical applications, we define the architecture of a SaaS Cloud system, which is based on the aforementioned engine. The resulting architecture allows devices with totally different and heterogeneous hardware and software characteristics to interact among them, so that these differences are almost completely transparent to the end-user. A Cloud-based solution for lossless dynamic and adaptive compression of 3D medical images.Management of such data may be considered as an atypical Big Data problem.It provides SaaS services based on an elastic and on-demand peer to peer overlay infrastructure.It also provides effective, secure and flexible access to healthcare resources that need to be managed by medical applications.Allows devices with totally different and heterogeneous hardware and software characteristics to interact among them.

Copyright Protection for Images on Mobile Devices

The rapid diffusion of mobile devices (telephones, smart phones, tablets, etc.) has brought new advanced features developed specifically for these devices. For example it is now possible to publish directly the pictures obtained by the integrated camera of a smart phone to our social network accounts, or to image hosting services, etc. It is therefore important to have tools on the mobile devices that can prove the ownership of the pictures and to use them before publishing the images. A common approach to guarantee ownership of digital images or videos is to use Digital Watermarking techniques. We have developed a tool for portable devices based on Android OS, that permits to add a visible or invisible watermark to images. In this paper we discuss this tool, and we experimentally prove its robustness.

A Secure Low Complexity Approach for Compression and Transmission of 3-D Medical Images

Digital images play an important role in a wide range of medical applications. Several widespread technologies for digital imaging, such as Computed Tomography (CT), Magnetic Resonance (MR), etc., produce three-dimensional images. Data compression, is thus essential to reduce the volume of such images, permitting their efficient storing along with the improvement of the relative transmission time through Internet or any other ad-hoc systems, like Picture Archiving and Communication System (PACS), Tele-radiology, etc.. Since these images are often stored in system particularly vulnerable from the point of view of security, especially because they contain sensitive data, it is necessary to provide such images with a mechanism which ensures at least security against message forgery. In fact, an attack can be made by altering a medical image, and consequently, may alter the relative diagnosis. The purpose of this work is twofold, first we propose a low complexity approach for the compression of 3-D medical images, then, in order to limit the above defined potential attack, we proposed en efficient method to insert within each image an invisible digital watermark, during the compression process. In this way, we define a hybrid approach that handles simultaneously and efficiently both the compression that the security of three-dimensional images. We validate the proposed approach by showing test results.

Visualization, Band Ordering and Compression of Hyperspectral Images

Air-borne and space-borne acquired hyperspectral images are used to recognize objects and to classify materials on the surface of the earth. The state of the art compressor for lossless compression of hyperspectral images is the Spectral oriented Least SQuares (SLSQ) compressor (see [1–7]). In this paper we discuss hyperspectral image compression: we show how to visualize each band of a hyperspectral image and how this visualization suggests that an appropriate band ordering can lead to improvements in the compression process. In particular, we consider two important distance measures for band ordering: Pearson’s Correlation and Bhattacharyya distance, and report on experimental results achieved by a Java-based implementation of SLSQ.

A Secure Distributed Video Surveillance System Based on Portable Devices

In this work a distributed video surveillance system based on a Client-Server architecture is presented. The proposed system is accessible from portable devices such as tablets, smartphones, etc. In a typical real-world scenario, for example in homeland security, it is useful to have portable devices that can receive in real-time a frame or a sequence of frames coming from a selected camera to prevent or to detect attacks (i.e. terrorist attacks, etc.). In the proposed system, a portable device knows only the address of the server (repository), and the repository sends to the portable device the list of the clients (nodes) which are connected with one or more cameras. When the portable device obtains the list of the nodes, it connects directly to a specific node and requests the images of its connected cameras. The whole system provides secure communication channel between all its components. The security of both the node-repository and the repository-portable devices communications is guaranteed by using a secure connection. The security of the node-portable devices interconnection is provided by a digital invisible watermarking algorithm that affects each image before sending it from the node to the portable devices. Each portable device can extract the watermark and verify the identity of the node.

A collaborative clinical analysis service based on theory of evidence, fuzzy linguistic sets and prospect theory and its application to craniofacial disorders in infants

Abstract Nowadays, it is more and more important to diagnose several kinds of pathologies at their early stage, in order to take the necessary countermeasures before having permanent consequences. Unfortunately, though many pathologies are widespread, there does not exist a unique standardized reference or Gold Standard according to which it is possible to evaluate the patients, mainly when the pathology is in the early stages or is not very noticeable, and the doctor is not sufficiently expert in the problem domain. In this work, we deal with this problem, by envisioning new healthcare services supporting a collaborative clinical analysis of symptoms collected from the patients and forwarded to a group of experts, which are geographically distributed. The experts return back their assessment and diagnosis and the system combines these by means of the Theory of the Evidence, in order to provide a single response. The above services can be easily implemented on top of state-of-the-art distributed computing facilities such as Grids or Clouds, providing a connected environment for medical data distributed over different sites and allowing medical experts to collaborate without being co-located, thereby providing transparent access to data and computing resources. Additionally, such services can provide feedbacks to each expert, in order to improve its own knowledge and experience in the case of divergence between the expert response and the global combined diagnosis in recognizing and classifying the received symptomatic indexes from the patient. We have considered the craniofacial pathologies in infant population as a practical example for better explaining the proposed solution.

Format-Independent Protection of DNA Microarray Images

DNA microarray images are an essential research tool in several fields, such as, forensic biology, single nucleotide polymorphism, functional analysis of genes, DNA sequencing, diagnosis of hereditary and infectious diseases, etc. Consequently, these images can be exchanged between different entities with different roles and belonging to different application domains. It is easy to note that for the reasons described above, the protection of such images, mainly in terms of authenticity and integrity is a main concern, especially in the so called multi-domain environments. It is important to remark that the use of conventional techniques for data protection, such as encryption or the adding of metadata into the image header is doomed to fail. Indeed, decrypted content may be subject to unauthorized use or manipulation, and metadata may be lost during the image processing, such as a format conversion. Therefore, in order to ensure the requirements of authenticity and integrity, it is necessary to use a security mechanism specific for DNA microarray images, which is as close as possible to such images. For this reason, in this paper we introduce a watermarking scheme explicitly designed for this kind of images. In particular, such a scheme belongs to the category of invisible reversible fragile watermarking. Finally, the validity of this scheme has been assessed through testing.

Adaptive Vector Quantization for Lossy Compression of Image Sequences

In this work, we present a scheme for the lossy compression of image sequences, based on the Adaptive Vector Quantization (AVQ) algorithm. The AVQ algorithm is a lossy compression algorithm for grayscale images, which processes the input data in a single-pass, by using the properties of the vector quantization to approximate data. First, we review the key aspects of the AVQ algorithm and, subsequently, we outline the basic concepts and the design choices behind the proposed scheme. Finally, we report the experimental results, which highlight an improvement in compression performances when our scheme is compared with the AVQ algorithm.

One-pass lossless data hiding and compression of remote sensing data

Abstract The information obtained by means of spectral remote sensing (i.e., the hyperspectral images) are involved in several real-life scenarios and applications. Historical research, monitoring of environmental hazards, forensics and counter-terrorism are some examples of contexts in which the hyperspectral data play an important role. In many contexts, the hyperspectral images could also play sensitive roles (e.g., in military applications, etc.) and are generally exchanged among several entities, in order to carry out different tasks on them. Therefore, it is important to guarantee their protection. A meaningful choice is the protection through data hiding techniques. In fact, by means of reversible data hiding techniques, the imaging data become a sort of information carrier and can be used for delivering other important data that can be used, for instance, to check the integrity of the original imaging data. In this paper, we introduce a one-pass framework that is able to perform the lossless data hiding and the lossless compression of the marked stream, at the same time, by exploiting the capabilities of the predictive paradigm. Substantially, in a single pass, a marked and compressed stego image is obtained, which can be exactly restored by the receiver: by decompressing and reversibly reconstructing the original unaltered image. In addition, our framework also permits to perform only the decompression (without the extraction of the hidden information). In this manner, the resulting stego (marked) hyperspectral image, could be used for several purposes, in which it is not necessary to extract the original data and an acceptable grade of degradation is tolerated. We also implement a proof-of-concept of the proposed framework to assess the effectiveness of our contribution. Finally, we report the achieved experimental results, which outperform other similar approaches.

A Collaborative Decision-Support System for Secure Analysis of Cranial Disorders

Nowadays cranial disorders are one of the most common type of infants diseases. However, despite such disorders are widespread, in the state of the art does not exist a system which allows the doctors to share their knowledge in order to support their clinical decisions. This is mainly caused by the different national laws governing the access to information concerning minors. However, it is important to emphasize that such limited access during the decision-making process, is one of the main causes of medical errors, and clearly can lead to serious consequences for the patients health. In order to overcome the above defined issues, we first propose an engine for the automatic analysis of the patients cranial data, which allows to detect end quantify eventual disorders. In addition, we propose a collaborative decision support system to enable the secure storing, sharing and analysis of clinically relevant information, which allows to not violate the patients privacy. In particular, the collaborative system we propose, enables the doctors to rely on a previously acquired base of knowledge, as well as on the opinions of other doctors, during all the phases necessary to carry out the analysis and the clinical evaluation of the patient under examination. The proposed collaborative system can represent a valuable and effective aid for the doctors in charge of such disorders, even allowing to overcome the limitations imposed by national law.