University of Milano-Bicocca Showcase

Raw data acquired with a Bruker IRIS hyperspectral scanner. Spectral reflectance data between 400 and 2500 nm. X-ray fluorescence data.

Accurate building extraction from high-resolution aerial imagery is essential for numerous applications in remote sensing, urban planning, and disaster management. While AI-based methods enable fast, scalable, and cost-effective segmentation of building footprints, their development is often limited by the scarce availability of large-scale, geographically diverse datasets with reliable pixel-level annotations. In this work, we present SegFVG, a large-scale, high-resolution, and geographically diverse dataset for building segmentation, focused on the Friuli Venezia Giulia region in northeastern Italy. The dataset includes over 15,000 true orthophoto aerial image tiles, each of size 2000 × 2000 pixels with a ground sampling distance of 0.1 meters, paired with precise pixel-level building segmentation masks. Covering approximately 616 square kilometers, SegFVG captures a broad spectrum of urban, suburban, and rural settings across varied landscapes, including mountainous, flat, and coastal areas. Alongside the dataset, we provide benchmark results using several deep learning models. These support the usability of SegFVG for the development of accurate segmentation models and serve as a baseline to accelerate future research in building segmentation.

Intestinal microbiota and gut-born T helper-17 (Th17) lymphocytes may act as drivers of smoldering multiple myeloma (SMM) to MM evolution. We demonstrate here that administering the human commensal Prevotella melaninogenica to transgenic Vk*MYC mice affected by Early-MM, mimicking human SMM, significantly delayed evolution to full-blown MM. Mechanistically, P. melaninogenica increased the production of short-chain fatty acids (SCFAs), preventing skew of dendritic cells towards a pro-Th17 phenotype and accumulation of Th17 cells in the bone marrow of treated mice. P. melaninogenica or butyrate synergized with anti-PD-L1 antibodies by restraining Th17 cell expansion while unleashing immune checkpoint blockade (ICB)-induced effector CD8+ T cells. P. melaninogenica also attenuated IL-17-mediated skin lesions that mimicked ICB-induced immune-related adverse events. Thus, modulation of the gut microbiota or SCFAs administration with or without ICB might represent treatment options for patients affected by plasma cell dyscrasias and other hematologic or solid tumors where IL-17 acts as driving force. The present dataset contains: - 1H NMR spectra acquired on control and growth media for PM and PH; - CMP and .exp files created for metabolite quantification in MestreNova software by SMA analysis.

The data reports experiments conducted with optcial and microwave assisted detection of magnetic field polarity for the determination of the oprational sensitivity and experimetnal parameters optimization.

An elite professional pianist executed a 30-minute, uninterrupted performance of seven pieces on a Yamaha P-225B digital piano in an anechoic chamber, employing the default soundbank. The repertoire featured Contrapunctus I (BWV 1080) by Bach and an excerpt from Chopin’s Ballade No. 1, Op. 23, performed from memory based on urtext editions. Key-press onset events were annotated per hand to quantify note and gesture counts, excluding legato transitions without discrete attacks. Both performances exhibited a convergent low-frequency periodicity in beat-level timing variability (tactus imprecision), oscillating at approximately 0.36 Hz. This slow temporal modulation aligns with the delta-band range of neural oscillations and may reflect a shared endogenous timing scaffold, plausibly motor in origin, underlying expressive control in skilled performance.

These data are part of a systematic study on the localized and non-localized etch rates of silicon nanopillars produced using Metal-Assisted Chemical Etching (MACE) for thermoelectric applications. Specifically, they represent a collection of etch rates as a function of doping level for both p-type and n-type silicon, using both 1-pot and 2-pot MACE processes. Data are available in both .csv and .xlsx format.

Experimental data collected from orichalcum chips. X-ray fluorescence (XRF), diffuse reflectance (DR) spectroscopy and combined XRF and VIS-SWIR reflectance maps

This dataset has been used in the manuscript titled “Snow Surface Parameters from PRISMA Data in the Alps and Their Implications for Radiative Forcing” and provides snow property parameters retrieved from PRISMA hyperspectral imagery acquired over Plateau Rosa (Aosta Valley, European Alps) following a Saharan dust deposition event in July 2024. The parameters include mineral dust concentration, snow grain size, broadband albedo, and instantaneous radiative forcing. These were estimated using a supervised learning approach based on the K-Nearest Neighbors (KNN) algorithm, which was applied to match observed spectra with a reference library generated using the BioSNICAR radiative transfer model. The dataset also includes Asymptotic Radiative Transfer (S-ART) model (S-ART) outputs and in situ measurements collected during a dedicated field campaign. These data support the investigation of the effects of light-absorbing impurities on snow radiative properties, contributing to the improvement of climate and hydrological modeling in high-altitude environments.

Supplementary data and underlying data supporting the finding described in "Production and carbon footprint of microbial oil from waste lemon peel extract". Abstract Background The agricultural sector is one of the leading producers of agro-industrial solid organic waste. This waste is mainly disposed of by incineration or landfilled, representing a huge loss of potential resources, which could be used for the production of high-value chemicals. In this study, a fermentation process for the production of microbial oil from waste lemon extract (LE), an aqueous side stream deriving from waste lemon peel and pulp processing, was developed and assessed for impacts. Microbial oil can have many and diverse applications, from plasticisers in plastic and rubber compounds to moisturizers in cosmetic formulation. Methods and results Characterization of LE revealed that its autoclaving process is effective for increasing the concentration of readily available glucose and fructose, reaching 28.77 ± 0.08 g L-1 and 25.68 ± 0.27 g L-1. Nitrogen content was measured too, revealing a C/N ratio of 85, optimal for triggering lipid accumulation in the selected microbial cell factory. Therefore, the oleaginous yeast Cutaneotrichosporon oleaginosum was cultivated in an unmodified LE-based medium in 2 L bioreactors, resulting in a lipid accumulation of 0.47 ± 0.08 goil gCDW-1. Finally, a new lipid extraction method using green solvents was developed, which allowed to extract and purify 11.29 g of oils, corresponding to 35% of the cell dry weight. The carbon footprint of this laboratory-scale production was estimated to be 71 - 434 kgCO2eq kg-1 microbial oil, with electricity consumption of the fermentation step as the main factor. Simulation of the process in a 300L fermenter suggests that the electricity consumption, and therefore the overall impact, can be drastically reduced with scale-up. Conclusions The proposed process is promising in terms of production and has the advantage of not being in competition with edible resources and land use. However, the microbial oil yield and the extraction process must be optimized to make the process sustainable.

XRF spectra acquired with IRIS