Pinto Almeida, Carlos

Preferred name

Main Affiliation

Facultad de Ingenierías

Ambato

Now showing 1 - 7 of 7

Some of the metrics are blocked by your
consent settings
Artificial intelligence for resilient transportation infrastructure: the first Ecuadorian pavement damage dataset and an assessment of domain transfer across international road networks
(Frontiers Media SA, 2026-08-26)
Zambrano-Salazar, Leonardo
;
López, Enma Gamboa
;
Pinto Almeida, Carlos
Background Pavement-condition monitoring is a cornerstone of resilient and sustainable transportation infrastructure, yet existing public datasets for automatic damage detection are dominated by data from high-income countries. Deep-learning models trained on those data often transfer poorly to Latin American road networks, where deterioration patterns, materials, and maintenance regimes differ. No publicly available pavement damage dataset for Ecuador has been reported to date, to the best of our knowledge. Methods We present the Ecuadorian Pavement Distress Dataset (EcuPDD), 1,295 images annotated under ASTM D6433 into six damage classes (4,559 annotated instances), collected across seven cities in the three geographic regions of Ecuador (Sierra, Coast, and Amazon). The dataset is released under CC BY 4.0 on Zenodo (DOI: 10.5281/zenodo.20454801 ). On this base we trained five YOLOv8n variants progressively fused with three open international datasets—RDD2022 Japan, RDD2022 China_MotorBike, and McMaster Pavement Crack Detection—for a total corpus of up to 15,941 images (of which roughly 80% were used for training and 20% for validation, following the splits described in Section 3). Each version was trained five times with independent random seeds (25 runs in total), and results are reported as mean ± standard deviation. We evaluated each version on its mixed validation set and, separately, on a fixed 259-image Ecuadorian held-out subset. We complemented the standard mAP50/ mAP50-95 metrics with Wilson 95% confidence intervals on per-class detection rates, a Gini coefficient for class imbalance, and a Domain Gap Index (DGI) that we define as 1 − (mAP50Ecuador/ mAP50general), summarising the divergence between the general and local performance of each model. Results Adding international data lifted the mixed-validation mAP50 from 14.65% (V1, Ecuador only) to 57.66% (V5, Ecuador + McMaster + China_MotorBike), an absolute gain of 43.01 percentage points. However, when evaluated only on the 259 Ecuadorian held-out images, mAP50 stayed between 14.65% and 15.90% across all five versions—a swing of 1.25 pp despite the inclusion of up to 14,646 additional images. The Wilson 95% confidence intervals on those Ecuadorian-only mAP50 values overlap heavily for all versions, indicating substantial overlap and no clear evidence of separation among versions at this sample size. The Domain Gap Index reaches 0.733 for V5 and 0.713 for V2, while remaining ∼0 for V1 and V3. Patching, despite contributing only 2.3% of the instances, reached a Wilson-CI-supported detection rate of 0.78 [0.73, 0.82] thanks to its strong visual distinctiveness when paired with the McMaster dataset. Conclusion International datasets improve a YOLOv8n detector globally but do not, by themselves, close the domain gap to Ecuadorian roads. The bottleneck is the volume and representativeness of local annotated data. EcuPDD provides the first openly licensed Ecuadorian benchmark for this task and an empirical calibration point that researchers in Latin America and other under-represented regions can use to plan dataset-construction and model-development efforts aligned with the resilience and sustainability goals of the AI-for-infrastructure agenda.
Some of the metrics are blocked by your
consent settings
AI-ready hydraulic risk prioritization of deteriorated Andean concrete channels
(Frontiers Media SA, 2026-09-01)
Núñez, Mayra Alejandra
;
Núñez, Samanda Monserrath
;
Guevara, Arlette Dayana
;
Pinto Almeida, Carlos
Introduction Open concrete channels in Andean watersheds convey irrigation water and urban stormwater, and their deterioration raises Manning’s roughness coefficient (n), alters specific energy (E) and consumes freeboard. Managers know that a channel is degraded but lack a reproducible procedure for turning that observation into a defensible intervention priority. Methods We characterise ten consecutive reaches (219 m) of a rectangular cyclopean-concrete channel in Ambato, Ecuador (b = 0.645 m; Q = 0.21 m 3 /s; s 0 ≈ 0.03). Each reach was scored on a 1–5 Structural Damage Index and mapped to a Manning coefficient through an additive decomposition reported reach by reach. The coefficients were propagated through the Manning–Strickler equation, solved simultaneously with continuity, to obtain normal depth, velocity, Froude number, specific energy and freeboard. Because these quantities are deterministic functions of the assigned n, they are reported as transfer curves rather than as statistical relationships; analysis is confined to the observations themselves and to the propagation of observer uncertainty. Results Assigned coefficients ranged from 0.016 to 0.034, that is 23%–162% above the design value of 0.013. Departure of specific energy from the design condition ranged from 14.5% to 30.9%, and freeboard fell from 0.277 m to 0.160 m in the most degraded reach, a loss of 42% of the original margin. The transfer curve is not monotonic: the departure peaks near n ≈ 0.030 and declines beyond it, because specific energy passes through its minimum at the critical regime, reached here at n ≈ 0.0304. A one-class error in visual scoring displaces the departure by 8.6 percentage points on average and would change the intervention tier of six of the ten reaches; the two requiring immediate intervention are not among them. Conclusion Structured visual assessment combined with transparent hydraulic propagation yields a traceable intervention hierarchy reproducible with a camera, a total station and an open-source solver. The hydraulic values are model-derived estimates conditional on the assigned coefficients, not measurements; calibration against in situ measurement remains necessary and generalisation beyond this channel is not demonstrated. The protocol, imagery and scripts are released as an openly licensed dataset intended as ground truth for future classifiers.
Some of the metrics are blocked by your
consent settings
Effect of replacing mixing water with milk on the compressive strength of concrete at different curing ages
(Frontiers Media SA, 2026-07-17)
Pinto Almeida, Carlos
;
Zambrano, Leonardo
;
Apraez, Jose
;
Gómez, Juan
;
Zurita, Marco
Introduction – The chemical composition of mixing water directly affects cement hydration and the strength development of concrete. Although alternative mixing liquids—such as treated wastewater, leachates, and wash water—have shown moderate strength reductions compatible with structural use, the behavior of concrete exposed to complex organic liquids remains largely uncharacterized. Methods – This study investigates how replacing mixing water with whole bovine milk, a medium rich in lactose, casein, and fat, affects cement hydration and compressive strength in conventional Portland concrete. Three mixtures with increasing replacement levels were produced: 0% (M0, control), 50% (M50), and 100% (M100). Twenty-seven cylindrical specimens (3 mixtures × 3 ages × 3 replicates) were tested under axial compression following ASTM C39/C39M at 15, 28, and 45 days. Peak strength, stress–strain curves, within-condition dispersion, and failure modes were recorded. Results – The control developed strengths of 14.39 ± 0.64, 16.47 ± 0.73, and 20.91 ± 0.81 MPa at 15, 28, and 45 days, whereas M50 and M100 remained below 1 MPa at all ages (M50: 0.61 ± 0.03, 0.67 ± 0.03, 0.90 ± 0.06 MPa; M100: 0.38 ± 0.02, 0.55 ± 0.02, 0.60 ± 0.03 MPa), a strength loss greater than 95% relative to the control. Coefficients of variation ranged from 3.9% to 6.1%, below the 10% threshold of ACI 214R. The modified mixtures also lost the ascending branch of the stress–strain curves and exhibited diffuse failure modes, without the longitudinal cracking typical of sound concrete. Conclusion – The magnitude of the strength loss (>95%) places milk in a category distinct from previously reported alternative mixing liquids. Rather than producing a degraded variant of conventional concrete, it generates a system in which the cementitious reaction is practically inhibited. This behavior, attributable to the combined action of lactose, casein, and fats on C3S hydration, redefines the operational limits of chemical compatibility between organic fluids and cementitious systems and provides a reference calibration point for predictive models of organic inhibition. Copyright © 2026 Pinto, Zambrano, Apraez, Gómez, Zurita and Buele.
Some of the metrics are blocked by your
consent settings
Mechanical Performance of Volcanic Ash Concrete Showing Modulus Reduction with Strength Retention
(2025)
Pinto Almeida, Carlos
;
Abril Camino, Andrés
;
Daniel Abril-Camino
This study aims to evaluate the mechanical behavior of concrete that incorporates 51.3% raw volcanic ash into its structure, focusing on its static elasticity modulus and compressive strength. Cylindrical concrete samples were prepared via the mix design commonly used in practice in Baños, Tungurahua, Ecuador. Three curing methods were applied: immersion, water spraying, and no curing. Compressive strength tests were conducted at 3, 7, 14, 21, and 28 days, whereas the static modulus was measured at 28 days following ASTM C469. Despite the high use of ash in the mixture, the mixtures achieved adequate compressive strengths for structural applications, reaching 28.05 MPa. However, a significant reduction in the static modulus was observed, with experimental values of approximately 7.06 GPa, whereas the value of 24.89 GPa was predicted by the equations given in ACI318. The use of raw volcanic ash in structural mixes requires modifications to deformation and stiffness calculations to ensure seismic performance, suggesting the need to review local regulations on traditional mixes. Based on the experimental data, an alternative empirical model, the VAM model, was proposed to better predict the elastic modulus of concrete with high volcanic ash content. The findings reveal a dual function of ash, acting as a pozzolanic material and as a low-density aggregate, highlighting the need to adjust the design equations when raw volcanic ash is used. This work contributes to the sustainable design of concrete mixtures in seismic regions.
10
Some of the metrics are blocked by your
consent settings
Digital Twins in Higher Education: Applications, Challenges, and Future Perspectives
(2026)
Pinto Almeida, Carlos
The Fourth Industrial Revolution has introduced emerging technologies that are reshaping educational models, among them digital twins, originally developed in the aerospace industry as virtual representations of real systems. Their integration into higher education now enables the simulation of complex learning environments, monitoring of student performance, and personalization of learning processes. However, their adoption within university settings remains incipient, fragmented, and lacking in standardization, which limits their transformative potential. This situation underscores the need to systematize dispersed knowledge on their educational use. This study presents a review with a descriptive and exploratory approach, based on peer-reviewed scientific articles selected from databases such as Scopus, IEEE Xplore, ScienceDirect, SpringerLink, and Web of Science, published between 2018 and 2024. The findings indicate that digital twins are primarily applied for simulations (68%), learning monitoring (52%), and educational personalization (44%). Reported outcomes highlight improvements not only in learner autonomy, motivation, and acquisition of practical skills, but also in conceptual understanding, problem-solving abilities, and collaboration among students in project-based environments. At the institutional level, benefits include greater efficiency in resource management, reduction of physical laboratory costs, and continuity of teaching in hybrid or remote learning contexts. Nevertheless, challenges remain, including limited connectivity, high implementation costs, insufficient teacher training, and ethical risks. Digital twins represent a strategic technology for transforming higher education into more flexible, personalized, and sustainable models, provided they are supported by inclusive policies and adequate investment. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.
3
Some of the metrics are blocked by your
consent settings
Lightweight deep learning for sustainable adobe construction: a block-level baseline aligned with circular economy principles and sustainable development goals
(Frontiers Media SA, 2026-07-08)
Pinto Almeida, Carlos
;
Jordan, Nancy
Introduction – The construction sector accounts for nearly 37% of global energy-related CO2 emissions, motivating renewed interest in low-carbon vernacular materials such as stabilized adobe. The integration of recycled inputs into adobe formulations advances circular-economy principles in the built environment, yet large-scale adoption is constrained by the absence of objective, scalable quality-control methods. Methods – We present a methodologically rigorous baseline for AI-driven visual quality control in stabilized adobe construction. A dataset of 330 images was constructed from 55 physical adobe blocks fabricated under controlled laboratory conditions with three circular-economy stabilization mixtures: recycled polyethylene terephthalate (1.5%), sugarcane bagasse fiber (8%), and Portland cement (10%). Each block was photographed from six standardized angles. To eliminate data leakage, we implemented strict block-level data splitting. A MobileNetV2 transfer learning pipeline with frozen ImageNet weights, aggressive data augmentation, L2 regularization, and balanced class weights was developed. Performance was characterized through accuracy with bootstrap 95% confidence intervals, per-class metrics, ROC AUC and Average Precision, five-fold block-level cross-validation, ablation study, and Grad-CAM interpretability. Results – The model achieved 80.95% test accuracy (95% CI: 69.0%–92.9%) on a held-out set of seven physical blocks (42 images). AUC = 0.858 and Average Precision = 0.944. Per-class F1-scores were 0.692 (sound) and 0.862 (defective). Block-level five-fold cross-validation confirmed stability at 70.0% (95% CI: 60.7%–79.2%). Ablation showed: simple CNN 71.43%, MobileNetV2 without augmentation 71.43%, complete pipeline 80.95% (+9.5 pp). Grad-CAM confirmed attention to physically meaningful surface features. The 9.24 MB model footprint enables offline deployment on commodity smartphones. Conclusion – This work establishes a methodologically rigorous baseline for AI-driven quality control in stabilized adobe construction. By demonstrating meaningful performance with modest datasets and edge-deployable architectures, this study supports broader adoption of circular-economy adobe construction and contributes to Sustainable Development Goals 9, 11, 12, and 13. Future work should expand the dataset, evaluate hierarchical classification, and assess field deployment. Copyright © 2026 Pinto Almeida and Jordán Buenaño. Conclusion This work establishes a methodologically rigorous baseline for AI-driven quality control in stabilized adobe construction. By demonstrating meaningful performance with modest datasets and edge-deployable architectures, this study supports broader adoption of circular-economy adobe construction and contributes to Sustainable Development Goals 9, 11, 12, and 13. Future work should expand the dataset, evaluate hierarchical classification, and assess field deployment.
Some of the metrics are blocked by your
consent settings
Exhaustive Differential Verification of Rule Execution in Building Compliance Software: A Method and Its Application to Seismic Provisions for Earthen Buildings
(MDPI AG, 2026-09-22)
Pinto Almeida, Carlos
;
Jordan, Nancy
;
Lucio, Diego Ricardo
;
Ambato-Jordán, Christopher
;
Mendoza, Alvaro Javier
Digital compliance tools increasingly stand between building regulations and the people who apply them, yet research on automated compliance checking has concentrated on whether a rule can be extracted from regulatory text rather than on whether the extracted rule is then executed correctly, so tools issuing compliance verdicts circulate without evidence of computational fidelity. This study proposes a verification method for the large class of provisions stating their requirements as algebraic inequalities over measurable building dimensions. The method combines division-free formalisation, cross-validation against an exact rational oracle, exhaustive enumeration of the discretised design space, differential comparison using asymmetric error classes, and a boundary coverage test suite. It was applied to the wall geometry provisions governing earthen buildings in Peru and Ecuador and to a mobile tool in free circulation. Exhaustive enumeration at the full input resolution of the tool, over 292,364,966,481 configurations, indicates that 7.93 percent of the baseline domain is admissible and that 72.1 percent of inadmissible configurations violate two or more requirements at once. The audit identified three execution defects; the rule sets disagree over 8.26 percent of the domain, and 55.9 percent of the verdicts issued lack normative support. A corrected version reproduced the reference verdict in all 32 device-executed test cases against 24 of 32 for the original, and agrees with the reference on every one of the 292,364,966,481 enumerated configurations. Applying the method to a second provision, the slenderness limit of EN 1996-1-1, reproduces the same defect signature in an unrelated code. Verification of rule execution is therefore inexpensive and necessary before compliance tools can be relied upon on site. The conditions under which the method applies are stated explicitly and are met by provisions in several codes, although transferability beyond the two cases examined remains to be demonstrated empirically. © 2026 by the authors.

Pinto Almeida, Carlos

Filters

Settings

Sort By

Results per page