Test Scenarios Research Articles

Forensic analysis of Scientific Linux image using commercial and opensource forensic tools

Depending on their needs and personal preferences, people choose to use different operating systems (OS) such as Windows, Linux, and Mac. The scientific Linux Operating System (SLOS) is designed to provide a stable, secure, and high-performance computing environment for scientific research and education in a steady, scalable, and extensible manner. When criminal activities are committed by suspects involving computers and the internet, it calls for digital forensics which involves the use of scientific procedures and tools to carry out the forensic investigation and analysis of digital evidence for legal and investigative purposes. Forensic investigators use commercial and opensource tools for analysis and gathering inculpatory and exculpatory pieces of evidence. This paper presents a comparative analysis of EnCase, FTK, Autopsy, bulk-extractor, and Scalpel for analyzing the Scientific Linux image. The test scenarios were designed to find out if the selected forensic tools can be appropriately used for investigating crimes committed using the SLOS. The test scenarios include extraction and analysis of operating system details, user accounts, web browsing history, and the recovery of deleted and shredded files and this paper compares and evaluates the capability of the tools in retrieving the evidence designed in the scenarios. This systematic comparison and evaluation results would assist digital forensics practitioners, researchers, and law enforcement agencies in making informed decisions regarding the selection of tools for Scientific Linux image forensics.

Modelling a DC Electric Railway System and Determining the Optimal Location of Wayside Energy Storage Systems for Enhancing Energy Efficiency and Energy Management

Global demand for fossil fuels is highly increasing, necessitating energy efficiency to be enhanced in transitioning to low-carbon energy systems. Electric railways are highly efficient in reducing the transportation demand for fossil fuels as they are lightweight and their energy demand can be fed by renewable energy resources. Further, the regenerative braking energy of decelerating trains can be fed to accelerating trains and stored in onboard energy storage systems (ESSs) and stationary ESSs. It is fundamental to model electric railways accurately before investigating approaches to enhancing their energy efficiency. However, electric railways are challenging to model as they are nonlinear, resulting from the rectifier substations, overvoltage protection circuits, and the unpredictability and uncertainty of the load according to the train position. There have been few studies that have examined the ESS location’s impact on improving the energy efficiency of electric railways while using specialised simulation tools in electric railways. However, no single study exists that has studied the location impact of stationary ESSs on the energy efficiency of electric railways while the trains are supported by onboard ESSs. Given these goals and challenges, the main objective of this work is to develop a model using commercial software used by industry practitioners. Further, the energy saving is aimed to be maximised using stationary ESSs installed in optimal locations while trains are supported by onboard ESSs. The model includes trains, onboard ESSs, rail tracks, passenger stations, stationary ESSs, and traction power systems involving power lines, connectors, switches, sectioning, and isolators. In this article, a test scenario is presented comprising two trains running on a 20 km with three passenger stations and two substations. The trains and track are modelled in OpenTrack simulation software (Version 1.9) while the power system is modelled in OpenPowerNet simulation software (Version 1.11). The two simulation tools are used in the railway industry and can produce realistic results by taking into account the entire electrical network structure. A stationary ESS is added on the wayside and moved in steps of 1 km to obtain the optimal location before investigating the impact of stationary ESSs on the performance and energy management of onboard ESSs. It is found that the energy saving when installing a stationary ESS at the optimal location is 56.05%, the peak-power reduction of Substation 1 is 4.37%, and the peak-power reduction of Substation 2 is 18.67%.

Successful use of the forced choice test for detecting concealment of semantic memory in criminal and intelligence investigations.

The current study examined the validity of the forced choice test (FCT) in a forensic scenario when used to detect concealment of semantic memory (SM-FCT). We also compared the SM-FCT validity to the FCT validity in the more commonly investigated episodic memory scenario (EM-FCT). In simulating a scenario of investigating suspected members of a terror organization, 277 students were asked to deceptively deny being enrolled in a college in which they do actually study. Results indicated that the SM-FCT's validity level was within the range of the EM-FCTs' validity levels. Theoretically, the results support a cognitive-based explanation for the FCT operation mechanism. Practically, they imply that FCT can be used in criminal or intelligence investigations of suspected members of terrorist or criminal organizations or suspected perpetrators of illegal acts or acts of terrorism, in which the incriminating evidence being sought is in the realm of designated semantic memory or knowledge.

Latch: Enabling large-scale automated testing on constrained systems

Testing is an essential part of the software development cycle. Unfortunately, testing on constrained devices is currently very challenging. First, the limited memory of constrained devices severely restricts the size of test suites. Second, the limited processing power causes test suites to execute slowly, preventing a fast feedback loop. Third, when the constrained device becomes unresponsive, it is impossible to distinguish between the test failing or taking very long, forcing the developer to work with timeouts. Unfortunately, timeouts can cause tests to be flaky, i.e., have unpredictable outcomes independent of code changes. Given these problems, most IoT developers rely on laborious manual testing.In this paper, we propose the novel testing framework Latch (Large-scale Automated Testing on Constrained Hardware) to overcome the three main challenges of running large test suites on constrained hardware, as well as automate manual testing scenarios through a novel testing methodology based on debugger-like operations—we call this new testing approach managed testing.The core idea of Latch is to enable testing on constrained devices without those devices maintaining the whole test suite in memory. Therefore, programmers script and run tests on a workstation which then step-wise instructs the constrained device to execute each test, thereby overcoming the memory constraints. Our testing framework further allows developers to mark tests as depending on other tests. This way, Latch can skip tests that depend on previously failing tests resulting in a faster feedback loop. Finally, Latch addresses the issue of timeouts and flaky tests by including an analysis mode that provides feedback on timeouts and the flakiness of tests.To illustrate the expressiveness of Latch, we present testing scenarios representing unit testing, integration testing, and end-to-end testing. We evaluate the performance of Latch by testing a virtual machine against the WebAssembly specification, with a large test suite consisting of 10,213 tests running on an ESP32 microcontroller. Our experience shows that the testing framework is expressive, reliable and reasonably fast, making it suitable to run large test suites on constrained devices. Furthermore, the debugger-like operations enable to closely mimic manual testing.

Multi-objective optimal power flow with wind–solar–tidal systems including UPFC using Adaptive Improved Flower Pollination Algorithm(AIFPA)

ABSTRACT Among the most significant non-linear challenges for power network design and smooth functioning of current modern updated power system networks is the optimum power flow (OPF) problem. Importance of the electrical system modeling has recently come to light due to the incremental use of energy from renewable sources in power systems networks. The goal is to use wind, solar and tidal sources to recreate the issue of OPF. In this work, Weibull, Lognormal, and also Gumbel probability distribution functions were applied to simulate the uncertainties of wind, photovoltaic, and tidal energy system. Additionally, by adding test scenarios of unpredictable wind, solar and tidal energy systems involving minimization of the cost function, loss of active power, voltage deviation, and increase stability of voltage. In accordance with the chosen thermal producing units, the solutions were evaluated using different locations using IEEE 30-bus systems of testing that incorporate renewable energy sources. The proposed power system planning problem was solved by multi-objective function where unified power flow controller are utilized as flexible AC transmission systems controllers via recently introduced optimization algorithms and the simulation outcomes of the aforementioned technique have been compared with Multi Objective Adaptive Guided Differential Evolution algorithms. The adaptive improved flower pollination algorithm (AIFPA) is a strong and reliable algorithm that is presented in this work. The AIFPA can efficiently deal with many kinds of high-complexity objective regions in multi-objective optimization situations. Utilizing an IEEE 30-bus test system, the suggested approaches’ performance was examined and evaluated for a range of objective functions. The simulation results obtained from the proposed algorithm were effective in finding the optimal solution compared to the results of the meta-heuristic algorithm and were reported in the literature.

A Red Team automated testing modeling and online planning method for post-penetration

Post-penetration red team automation testing effectively addresses the pain points of traditional manual red teams, including manpower, time costs, and the high level of professional knowledge required, thereby improving the efficiency and effectiveness of red team penetration testing. However, introducing automation technology into the red team testing domain still faces numerous technical challenges. These challenges include accurately simulating real attack environments, coordinating complex attack actions, and effectively resolving uncertainties during the attack process. These challenges remain critical issues that require urgent solutions. In this context, we propose a post-penetration-oriented automated red team penetration test modeling and planning approach. The objective of this approach is to automatically generate attack paths, coordinate attack behaviors, and adjust attack behaviors based on feedback, enabling attacks on real target networks through corresponding operations. We conducted analysis and performance testing on our solution, comparing it with other available planners. Our experimental results demonstrate the effectiveness of the proposed planner in achieving automated penetration testing. Compared to other available planners, ours can generate valid attack paths more quickly and exhibits excellent performance in planning effectiveness and quality. Furthermore, our planner possesses wide applicability across various penetration testing scenarios.

Real-virtual Point Cloud Augmentation Method for Test and Evaluation of Autonomous Weapon Systems

Autonomous weapon systems act according to artificial intelligence-based judgement based on recognition through various sensors. Test and evaluation for various scenarios is required depending on the characteristics that artificial intelligence-based judgement is made. As a part of this approach, this paper proposed a LiDAR point cloud augmentation method for mixed-reality based test and evaluation. The augmentation process is achieved by mixing real and virtual LiDAR signals based on the virtual LiDAR synchronized with the pose of the autonomous weapon system. For realistic augmentation of test and evaluation purposes, appropriate intensity values were inserted when generating a point cloud of a virtual object and its validity was verified. In addition, when mixing the generated point cloud of the virtual object with the real point cloud, the proposed method enhances realism by considering the occlusion phenomenon caused by the insertion of the virtual object.

Deception abilities emerged in large language models

Large language models (LLMs) are currently at the forefront of intertwining AI systems with human communication and everyday life. Thus, aligning them with human values is of great importance. However, given the steady increase in reasoning abilities, future LLMs are under suspicion of becoming able to deceive human operators and utilizing this ability to bypass monitoring efforts. As a prerequisite to this, LLMs need to possess a conceptual understanding of deception strategies. This study reveals that such strategies emerged in state-of-the-art LLMs, but were nonexistent in earlier LLMs. We conduct a series of experiments showing that state-of-the-art LLMs are able to understand and induce false beliefs in other agents, that their performance in complex deception scenarios can be amplified utilizing chain-of-thought reasoning, and that eliciting Machiavellianism in LLMs can trigger misaligned deceptive behavior. GPT-4, for instance, exhibits deceptive behavior in simple test scenarios 99.16% of the time (P < 0.001). In complex second-order deception test scenarios where the aim is to mislead someone who expects to be deceived, GPT-4 resorts to deceptive behavior 71.46% of the time (P < 0.001) when augmented with chain-of-thought reasoning. In sum, revealing hitherto unknown machine behavior in LLMs, our study contributes to the nascent field of machine psychology.

Development of an energy-autonomous eddy current sensor system for in-situ component monitoring

Energy-autonomous non-destructive testing (NDT) systems provide new opportunities for structural health monitoring. An NDT system that is energy-autonomous and transmits the recorded data wirelessly can be used in places that are not easily accessible for inspection. In addition, it can provide for continuous monitoring, which can improve the operational safety of various components and structures. This study presents a concept that enables energy-autonomous and wireless structural health monitoring using a splined shaft as a demonstrator. To monitor mechanical overloads, a material sensor is applied to critical areas of these components by a laser heat treatment. The structural change that occurs as a result of an overload can be monitored by the developed smart sensor system. This system consists of a material sensor, a low-cost eddy current sensor, an energy-efficient evaluation unit with an ultra-low-power microcontroller unit (MCU) and a Long-Range Wide Area Network (LoRaWAN) data transmission unit. Wireless data transmission via LoRaWAN provides access to the Industrial Internet of Things network, enabling the collection of NDT data from multiple smart sensors in a cloud and the processing of these data to assess the condition of the component from any location. It is shown that the MCU-based eddy current testing system is able to generate a similar signal quality for material characterisation compared to an industrial eddy current testing system, despite a much more compact design and a significantly higher energy efficiency. The energy autonomy of the smart sensor system is achieved by an energy harvesting system integrated into the splined shaft. In addition, three different testing scenarios are presented, which differ mainly in the scope of data processing and evaluation on the MCU, and thus also in the energy required.

RCDNet: An Interpretable Rain Convolutional Dictionary Network for Single Image Deraining.

As common weather, rain streaks adversely degrade the image quality and tend to negatively affect the performance of outdoor computer vision systems. Hence, removing rains from an image has become an important issue in the field. To handle such an ill-posed single image deraining task, in this article, we specifically build a novel deep architecture, called rain convolutional dictionary network (RCDNet), which embeds the intrinsic priors of rain streaks and has clear interpretability. In specific, we first establish a rain convolutional dictionary (RCD) model for representing rain streaks and utilize the proximal gradient descent technique to design an iterative algorithm only containing simple operators for solving the model. By unfolding it, we then build the RCDNet in which every network module has clear physical meanings and corresponds to each operation involved in the algorithm. This good interpretability greatly facilitates an easy visualization and analysis of what happens inside the network and why it works well in the inference process. Moreover, taking into account the domain gap issue in real scenarios, we further design a novel dynamic RCDNet, where the rain kernels can be dynamically inferred corresponding to input rainy images and then help shrink the space for rain layer estimation with few rain maps, so as to ensure a fine generalization performance in the inconsistent scenarios of rain types between training and testing data. By end-to-end training such an interpretable network, all involved rain kernels and proximal operators can be automatically extracted, faithfully characterizing the features of both rain and clean background layers and, thus, naturally leading to better deraining performance. Comprehensive experiments implemented on a series of representative synthetic and real datasets substantiate the superiority of our method, especially on its well generality to diverse testing scenarios and good interpretability for all its modules, compared with state-of-the-art single image derainers both visually and quantitatively. Code is available at https://github.com/hongwang01/DRCDNet.

Transforming clinical clerkship processes with integrated information systems

There are still many manual administrative processes involved in clinical clerkships. However, digital technologies and software solutions can significantly improve medical education. This paper aims to utilize extreme programming (XP) as a methodology for developing the information system. A total of 23 user stories have been planned for completion. The priority of each requirement stated in the user stories is determined using the MoSCoW prioritization technique. The results indicate that the software successfully passed all tests among the 23 black box testing scenarios. Developing an integrated information system to streamline clinical clerkships has been successfully achieved. The user experience questionnaire (UEQ) is utilized to gather the perspectives of hospital management and medical faculty members regarding the information system. The average scores for each category are as follows: 1.93 for attractiveness, 2.04 for perspicuity, 2.20 for efficiency, 2.06 for dependability, 2.14 for stimulation, and 1.85 for novelty. Therefore, it can be concluded that all existing user experience scales exhibit an outstanding degree of user satisfaction.

Implementasi Single Sign-On Menggunakan Protokol Openid Connect (OIDC) Pada Virtual Private Server (VPS)

Single Sign-On (SSO) is an authentication method that allows users to access multiple website services using one account and one login process. OpenID Connect (OIDC) is a protocol based on OAuth 2.0 and provides user identity information through JWT (JSON Web Token) tokens. Virtual Private Server (VPS) is a service that can be configured according to user needs. Internet users have many accounts needed to access various online services such as email, social media, online shopping, and many more. Each online service usually has a different username and password. Users often find it difficult to remember each password needed. In addition, using the same password on each online service is also not recommended because it can increase security risks. Based on this problem, an SSO system is needed that is useful for helping users in logging in. The SSO system will be tested with two Dummy Students and Dummy Elearning clients using Black Box Testing with the Equivalence Partitioning technique. From the test results, it was obtained that the SSO system showed that 83% of the test scenarios were successfully carried out

Sampling-based test scenarios searching method for autonomous system robustness evaluation

The robustness evaluation for autonomous systems is essential before implementing them in safety-critical fields. Searching the challenging test scenarios by adaptive sampling algorithm in the transition regions between different performance modes, named performance boundary, becomes an efficient method to evaluate the robustness of systems. This paper proposes a novel model-independent nearest neighbors adaptive sampling algorithm considering gradient and distance, named GDNNAS. GDNNAS designs a special evaluation criterion to identify samples within boundary regions and a searching method to generate new samples. A novel elimination criterion is proposed to ensure the quality of samples and an adaptive stopping criterion is proposed to stop the sampling process adaptively. Considering the lack of unified data sets and metrics to assess adaptive sampling methods for searching the performance boundary, we proposed a comparison platform consisting of nine simulated systems under test (SSUT) as adaptive sampling benchmarks and four evaluation metrics to evaluate the adaptive sampling method quantitatively. We assess GDNNAS and two other adaptive sampling algorithms with the nine SSUTs. The experimental result proves that samples generated by GDNNAS are precisely located at and cover the performance boundary regions. In addition, we analyze the effectiveness of the comparison platform. Experimental result shows that the proposed SSUTs and evaluation metrics can comprehensively reflect the performance of the algorithm from various aspects and reveal the characteristics of the adaptive sampling algorithms. To verify the practicability of GDNNAS, a path planning system is used for testing GDNNAS, and the experiment result proves that GDNNAS performs outstandingly in practical system robustness evaluations.

Using cellular automata and multi-criteria evaluation to simulate the wildfire expansion in Prince George, British Co-lumbia

Wildfires pose a critical and ongoing challenge in British Columbia, Canada, threatening human life, property, and natural ecosystems. To understand and predict the behavior of such fires, our study employs Cellular Automata (CA), a mathematical model adept at simulating complex systems through grid-based cell interactions. This model, validated by prior research, incorporates a wind propagation rule that significantly enhances the prediction of wildfire spread in the direction of prevailing winds. Research centers on a wildfire event in Prince George, utilizing CA to simulate fire dynamics influenced by var-ious factors. The models strength lies in its ability to represent detailed local interactions and its flexibility in scenario testing, which is instrumental in understanding model uncer-tainties. By simulating different fire scenarios, the study aims to grasp the complexities and potential variables affecting wildfire behavior. The research provides a foundation for decision-makers to analyze and study wildfire events, leveraging a Multi-Criteria Evaluation (MCE) Model to assess the susceptibility of cells to fire. This comprehensive approach combines CA with MCE, offering a robust framework for simulating and manag-ing wildfire expansion in British Columbia.

Bayesian Safety Validation for Failure Probability Estimation of Black-Box Systems

Estimating the probability of failure is an important step in the certification of safety-critical systems. Efficient estimation methods are often needed due to the challenges posed by high-dimensional input spaces, risky test scenarios, and computationally expensive simulators. This work frames the problem of black-box safety validation as a Bayesian optimization problem and introduces a method that iteratively fits a probabilistic surrogate model to efficiently predict failures. The algorithm is designed to search for failures, compute the most-likely failure, and estimate the failure probability over an operating domain using importance sampling. We introduce three acquisition functions that aim to reduce uncertainty by covering the design space, optimize the analytically derived failure boundaries, and sample the predicted failure regions. Results show this Bayesian safety validation approach provides a more accurate estimate of failure probability with orders of magnitude fewer samples and performs well across various safety validation metrics. We demonstrate this approach on three test problems, a stochastic decision-making system, and a neural-network-based runway detection system. This work is open-sourced (https://github.com/sisl/BayesianSafetyValidation.jl) and is currently being used to supplement the FAA certification process of the machine learning components for an autonomous cargo aircraft.

Development of a dynamic traffic microsimulator for on-demand transit operations within an integrated modelling system

ABSTRACT This study develops an intelligent dynamic agent-based microsimulation (iDAMS) module for on-demand transit (ODT) operations within an integrated transport, land-use and energy (iTLE) model. A real-time optimization component within the iDAMS is formulated by the utilization of travelling salesman problem and simulated annealing metaheuristics that perform dynamic passenger-vehicle assignment. It simulates ODT operations for meeting the 24-hour auto-trip demand of Halifax, Canada, to compare the performance of the proposed system with personal cars (PCs). The optimization objectives are to determine the optimal fleet size and seat capacity that satisfies maximum trip requests while minimizing waiting time, travel time and vehicle kilometres travelled (VKT). Simulation results indicate that the ODT system can deliver service similar to PCs in Halifax while decreasing cost and emissions (13% reduction in VKT). The tools developed in this research will provide transit planners ability to conduct ODT scenario simulation and test system performance in real time.

MULTI-BLOCK: A novel ML-based intrusion detection framework for SDN-enabled IoT networks using new pyramidal structure

The ever-expanding Internet of Things (IoT) landscape faces significant security challenges due to limitations in traffic monitoring, device heterogeneity, and weak security practices, leaving networks vulnerable to multi-target and coordinated attacks like large-scale botnets and Distributed Denial-of-Service (DDoS). Existing intrusion detection systems can be computationally expensive and resource-intensive. This can be problematic for resource-constrained IoT devices with limited processing power, memory, and battery life. This paper proposes MULTI-BLOCK, a novel, multi-module framework that leverages machine learning, stateful P4 processing, and a Software-Defined Networking (SDN)-based multi-controller architecture. MULTI-BLOCK tackles critical management tasks within IoT networks, including synchronized communication, traffic monitoring, intrusion detection, and attack mitigation. This comprehensive framework comprises four modules. The first, the proposed pyramidal conceptually decentralized multi-controller structure (PCDMCS), introduces Decentralized Control Interfaces (DCIs) for real-time threat identification through a Decentralized Warning Conduit (DWC). The second module provides comprehensive network monitoring using P4-enabled 24-state tables. The third module leverages 30 new P4-extracted/calculated features and the Enhanced Weighted Ensemble Algorithm (EWEA) for enhanced anomaly detection. The final module presents a novel mitigation approach with 22 steps distributed across multiple controllers for scalability. Extensive evaluation using established IoT datasets (X-IIoTID, TON_IoT, and Edge-IIoTset) and diverse test scenarios (including single-victim and multi-victim attacks) demonstrates MULTI-BLOCK's exceptional performance, achieving high accuracy (99.75 %), precision (99.32 %), F1-score (99.53 %), recall (99.67 %), specificity (99.60 %), low false positive rates (FPR) (0.40 %), and low Average Detection Time (ADT) (2.11 ms). By offering a robust and adaptable solution against evolving threats, MULTI-BLOCK represents a significant advancement in IoT network security.

Secure authentication of identity information for quantum resistant IoT devices based on R-LWE and QIBE

ABSTRACT In order to improve the security of Internet of Things device authentication, a security framework is built in this paper by combining the ring-learning band error problem with quantum indistinguishable bilinear encryption. The research method meticulously designs different test scenarios and conducts multiple rounds of testing to evaluate the accuracy of the system authentication. The results indicated that the authentication accuracy was improved to 93.3% on average after using the proposed algorithm. The authentication speed is significantly improved with a shorter operation cycle. After parameter optimization, repeated tests indicated that the accuracy rate remained stable at over 90%, and the accuracy rate reached 100% in specific phases. Combining the ring learning with error problem technique with quantum indistinguishable bilinear encryption greatly improves the authentication security of Internet of Things devices against quantum computing threats. This discovery provides new strategies for Internet of Things security against quantum attacks and has far-reaching theoretical and practical implications for protecting critical infrastructure.

Improvement of Computer Adaptive Multistage Testing Algorithm Based on Adaptive Genetic Algorithm

Multistage testing (MST) is a portion of computational adaptive testing that adapts assessment structure at the sublevel rather than the component level. The goal of the MST algorithm is to identify bugs in computer programming, and there is a significant cost to utilising MST due to its decreased versatility during software development and maintenance. The efficiency of most algorithms drastically reduces for adaptive MST with complex feasible regions, while some modern algorithms function well while tackling computerised MST with a basic practicable range. The study offers an automated Adaptive Multistage Testing algorithm based on Adaptive Genetic Algorithm (AMST-AGA) for optimisation and scalability problems, in which constraints are successively introduced and dealt with at various evolutionary phases. In this paper, many test cases will aid in finding bugs and meeting completeness goals. Each time test cases are created, these testing scenarios must continue to pass.

Repository of MRI-derived models of the breast with single and multiple benign and malignant tumors for microwave imaging research.

The diagnosis of breast cancer through MicroWave Imaging (MWI) technology has been extensively researched over the past few decades. However, continuous improvements to systems are needed to achieve clinical viability. To this end, the numerical models employed in simulation studies need to be diversified, anatomically accurate, and also representative of the cases in clinical settings. Hence, we have created the first open-access repository of 3D anatomically accurate numerical models of the breast, derived from 3.0T Magnetic Resonance Images (MRI) of benign breast disease and breast cancer patients. The models include normal breast tissues (fat, fibroglandular, skin, and muscle tissues), and benign and cancerous breast tumors. The repository contains easily reconfigurable models which can be tumor-free or contain single or multiple tumors, allowing complex and realistic test scenarios needed for feasibility and performance assessment of MWI devices prior to experimental and clinical testing. It also includes an executable file which enables researchers to generate models incorporating the dielectric properties of breast tissues at a chosen frequency ranging from 3 to 10 GHz, thereby ensuring compatibility with a wide spectrum of research requirements and stages of development for any breast MWI prototype system. Currently, our dataset comprises MRI scans of 55 patients, but new exams will be continuously added.