Florida Polytechnic University

This paper introduces VERA (Vision Expert Real Analysis), a privacy-supporting cyber-physical framework designed for real-time data collection and visual analysis in healthcare environments. VERA limits exposure to identifiable RGB content by ensuring that annotators interact only with non-identifiable edge-based representations, while original images remain encrypted at rest using AES-CFB, with integrity verification performed before in-memory decryption. The system integrates edge-based obfuscation, secure annotation, in-memory decryption, and dynamic data augmentation to train YOLO-based person detection models without compromising patient privacy. Experimental results on a curated COCO subset show that VERA enables effective person detection, improving mean Average Precision (mAP) from an intentionally minimal baseline of 0.61 percent to 99.94 percent after full training and augmentation. This baseline is used solely to illustrate the contribution of the secure data preparation pipeline and is not intended to represent a fully optimized YOLO configuration. The results demonstrate that privacy-supportive workflows can maintain strong model performance while aligning with data protection practices common in regulated environments. Although this work focuses on person detection as a foundational stage, the VERA architecture is designed to support future extensions toward privacy-preserving Human Activity Recognition (HAR) tasks in clinical and assisted-living settings.

The nickel-based superalloy Inconel 718 (IN718) is commonly used in the aerospace industry in the fabrication of gas engine turbine blades, which undergo a combination of cyclic axial and torsional loadings coupled with extreme environments during service. Given the geometric complexity of these blades, additive manufacturing (AM), and more specifically, the material extrusion additive manufacturing (MEAM) process offers a potential cost-effective and easily accessible solution to these manufacturing challenges. The shortfall of this technique, however, is the potential risk of manufacturing defects introduced during part production, such as warping, surface texture, and shrinkage that could negatively impact the fatigue life and performance of these specimens. This study aims to investigate the torsional fatigue performance of MEAM IN718 as-built green parts of X- and Z-build orientation, and the role of processing parameters and manufacturing defects in the failure of these components, with on-going work exploring the role of debinding and sintering stages on performance. This knowledge can provide insight into the viability of MEAM for use within the aerospace industry.

Repackaged Android applications pose a significant threat to mobile ecosystems, acting as common vectors for malware distribution and intellectual property infringement. Addressing the challenges of existing repackaging detection methods—such as scalability, reliance on app pairs, and high computational costs—this paper presents a novel hybrid approach that combines supervised learning and symptom discovery. We develop a lightweight feature extraction and analysis framework that leverages only 20 discriminative features, including inter-component communication (ICC) patterns, sensitive API usage, permission profiles, and a structural anomaly metric derived from string offset order. Our experiments, conducted on 8441 Android applications sourced from the RePack dataset, demonstrate the effectiveness of our approach, achieving a maximum F1 score of 85.9% and recall of 98.8% using Support Vector Machines—outperforming prior state-of-the-art models that utilized over 500 features. We also evaluate the standalone predictive power of AndroidSOO’s string offset order feature and highlight its value as a low-cost repackaging indicator. This work offers an accurate, efficient, and scalable alternative for automated detection of repackaged mobile applications in large-scale Android marketplaces.

Abstract Flow boiling continues to be an ever-growing and effective heat transfer process with applications in high-power electronics systems like servers. Flow boiling can be improved through the use of different surface geometries. This research investigated the geometries of micro-fin (diamond and straight) and mini-cavity (staggered and aligned with a secondary cavity) arrays in flow boiling. The micro fin specimens have a much higher surface area, leading to an increase in heat transfer efficiency. The cavities act as artificial bubble propagation sites, allowing for the boiling process to be more controlled. A baseline specimen with no protuberances or cavities is also tested for comparison. Constant flow rates of 0.1, 0.2, 0.3, and 0.4 L/min have been tested. The heating powers of 217.6 W, 340.2 W, and 500 W were kept constant during the tests. To determine the convective coefficient of the different micro fins and cavities samples, they were run until steady-state (SS) conditions. The experimental results showed that the fins array performed better than the diamond micro-fins and cavities. The flow boiling process on micro-fin surfaces is an effective mechanism for dissipating high heat inputs.

We present a conditional variational autoencoder (CVAE) that generates stellar spectra covering 4000 ≤ Teff ≤ 11,000 K, 2.0≤logg≤5.0 dex, −1.5≤[M/H]≤+1.5 dex, vsini≤300 km/s, ξt between 0 and 4 km/s, and for any instrumental resolving powers less than 115,000. The spectra can be calculated in the wavelength range 4450–5400 Å. Trained on a grid of SYNSPEC spectra, the network synthesizes a spectrum in around two orders of magnitude faster than line-by-line radiative transfer. We validate the CVAE on 104 test spectra unseen during training. Pixel-wise statistics yield a median absolute residual of <1.8×10−3 flux units with no wavelength-dependent bias. A residual error map across the parameters plane shows ⟨|ΔF|⟩<2×10−3 everywhere, and marginal diagnostics versus Teff, logg, vesini, ξt, and [M/H] reveal no relevant trends. These results demonstrate that the CVAE can serve as a drop-in, physics-aware surrogate for radiative transfer codes, enabling real-time forward modeling in stellar parameter inference and offering promising tools for spectra synthesis for large astrophysical data analysis.

Parkinson's disease (PD) is a progressive neurological disorder caused by the gradual death of neurons throughout the body. This neuron death particularly affects several key neurotransmitter-producing neurons, including dopaminergic, serotonergic, noradrenergic, and cholinergic neurons located in the substantia nigra pars compacta, raphe nuclei, locus coeruleus, and nucleus basalis, respectively. Dopaminergic neurons are most associated with PD, with the decrease in dopamine believed to be the primary driver behind the "triad" of parkinsonian motor symptoms: bradykinesia, rigidity, and tremor. However, theories for how dopamine depletion modifies human motor control do not yet link all three symptoms to this single dysfunction. This proposed model expands upon a tremor model based on unstable feedback to capture all three symptoms. Furthermore, the model implements high-level control parameters that capture the potential effects of other neurotransmitters on symptom expression. This new type of model captures the existence of PD subtypes based on the relative levels of neuron death in different brain regions. This knowledge could provide insight into how differing disease progression across brain regions may produce the varying symptoms experienced across PD patients.