Magda Alexandra Trujillo-Jiménez
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Magda Alexandra Trujillo-Jiménez has a total of 68 co-authors. Here are the top co-authors by number of shared publications:
- Claudio Delrieux
Departamento de Ingeniería Eléctrica y Computadoras, Instituto de Ciencias e Ingeniería de la Computación, Universidad Nacional del Sur-Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Bahía Blanca 8000, Argentina
5 shared publications - Rolando González-José
Centro Nacional Patagónico (CENPAT-CONICET)
5 shared publications - Anahí Ruderman
Instituto Patagónico de Ciencias Sociales y Humanas, CCT CENPAT-CONICET, Puerto Madryn, Argentina; Programa de Referencia y Biobanco Genómico de la Población Argentina, Secretaría de Planeamiento y Políticas, Ministerio de Ciencia, Tecnología e Innovación, Ciudad Autónoma de Buenos Aires, Argentina
4 shared publications - Luis Orlando Pérez
Instituto Patagónico de Ciencias Sociales y Humanas, CCT CENPAT-CONICET, Puerto Madryn, Argentina; Programa de Referencia y Biobanco Genómico de la Población Argentina, Secretaría de Planeamiento y Políticas, Ministerio de Ciencia, Tecnología e Innovación, Ciudad Autónoma de Buenos Aires, Argentina
4 shared publications - Pablo Navarro
Instituto Patagónico de Ciencias Sociales y Humanas CENPAT Puerto Madryn Argentina
4 shared publications - Soledad De Azevedo
Instituto Patagónico de Ciencias Sociales y Humanas, CCT CENPAT-CONICET, Puerto Madryn, Argentina; Programa de Referencia y Biobanco Genómico de la Población Argentina, Secretaría de Planeamiento y Políticas, Ministerio de Ciencia, Tecnología e Innovación, Ciudad Autónoma de Buenos Aires, Argentina
4 shared publications
Latest Publications
- 1 September 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Latin America Transactions
- Vol. 24 (9) , 857-868
Satellite imagery super-resolution using GANs and aerial images
- 1 September 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Latin America Transactions
- Vol. 24 (9) , 857-868
Satellite imagery super-resolution using GANs and aerial images
- Open Access
- 9 March 2026
- Published by Universidad Nacional de La Plata in SADIO Electronic Journal of Informatics and Operations Research
- Vol. 25 (1) , e098
Super-resolución de imágenes satelitales usando GAN: un esquema basado en entrenamiento con imágenes aéreas
Las imágenes satelitales suelen presentar limitaciones en su resolución espacial y, en muchos casos, altos costos de adquisición, lo que restringe su uso en aplicaciones como el monitoreo urbano, la gestión territorial y el estudio de fauna. Este trabajo propone un enfoque innovador que aprovecha imágenes aéreas de alta resolución para entrenar un modelo de super-resolución basado en Redes Generativas Adversarias. En particular, se adapta el modelo ESRGAN (Enhanced Super-Resolution Generative Adversarial Network), optimizando sus parámetros con el fin de mejorar su eficiencia computacional y reducir los tiempos de entrenamiento. El modelo entrenado con imágenes aéreas se evalúa posteriormente sobre recortes satelitales de baja resolución, analizando su desempeño en factores de escala x2 y x4 mediante métricas estructurales, perceptuales y cromáticas (SSIM-Y, MS-SSIM, LPIPS y CIEDE2000). Los resultados muestran mejoras visuales claras, con mayor nitidez, mejor definición de bordes y una recuperación coherente de estructuras urbanas y elementos del terreno. Cuantitativamente, la escala x2 alcanza los valores más altos, mientras que la escala x4 mantiene un rendimiento estable y útil para aplicaciones prácticas. Estos hallazgos demuestran la viabilidad de transferir la capacidad de super-resolución desde imágenes aéreas hacia imágenes satelitales, incluso bajo diferencias espectrales y geométricas entre dominios. En conjunto, este trabajo establece una base sólida para el desarrollo de modelos de super-resolución satelital de bajo costo y alto impacto, y abre futuras líneas de investigación orientadas a ampliar los datos de entrenamiento, incorporar técnicas de domain adaptation y explorar arquitecturas específicas para sensores satelitales.
- 9 March 2026
- Published by Universidad Nacional de La Plata in SADIO Electronic Journal of Informatics and Operations Research
- Vol. 25 (1) , e098
Open AccessSuper-resolución de imágenes satelitales usando GAN: un esquema basado en entrenamiento con imágenes aéreas
Las imágenes satelitales suelen presentar limitaciones en su resolución espacial y, en muchos casos, altos costos de adquisición, lo que restringe su uso en aplicaciones como el monitoreo urbano, la gestión territorial y el estudio de fauna. Este trabajo propone un enfoque innovador que aprovecha imágenes aéreas de alta resolución para entrenar un modelo de super-resolución basado en Redes Generativas Adversarias. En particular, se adapta el modelo ESRGAN (Enhanced Super-Resolution Generative Adversarial Network), optimizando sus parámetros con el fin de mejorar su eficiencia computacional y reducir los tiempos de entrenamiento. El modelo entrenado con imágenes aéreas se evalúa posteriormente sobre recortes satelitales de baja resolución, analizando su desempeño en factores de escala x2 y x4 mediante métricas estructurales, perceptuales y cromáticas (SSIM-Y, MS-SSIM, LPIPS y CIEDE2000). Los resultados muestran mejoras visuales claras, con mayor nitidez, mejor definición de bordes y una recuperación coherente de estructuras urbanas y elementos del terreno. Cuantitativamente, la escala x2 alcanza los valores más altos, mientras que la escala x4 mantiene un rendimiento estable y útil para aplicaciones prácticas. Estos hallazgos demuestran la viabilidad de transferir la capacidad de super-resolución desde imágenes aéreas hacia imágenes satelitales, incluso bajo diferencias espectrales y geométricas entre dominios. En conjunto, este trabajo establece una base sólida para el desarrollo de modelos de super-resolución satelital de bajo costo y alto impacto, y abre futuras líneas de investigación orientadas a ampliar los datos de entrenamiento, incorporar técnicas de domain adaptation y explorar arquitecturas específicas para sensores satelitales.
- Open Access
- 24 October 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-16
Genetic ancestry influences body shape and obesity risk in Latin American populations
Obesity is not simply a matter of excess weight. It also involves changes in structure and proportion in body morphology that can vary between populations and within individuals as they develop and age. Anthropometric measurements and their derived indices are widely used to study obesity. However, they present limitations to capture variations of fat distribution in the human body within a given population, and among different populations. Particularly, currently a problem in epidemiology is that cut-off points and health risk classifications based on anthropometric measures such as BMI, WHR or WHtR may not be equally valid for all population groups, especially when there are differences in genetic ancestry. Using data from \(\sim 7,000\) Latin American adults, we evaluated the accuracy of traditional indices across gradients of Native American, European, and African ancestry, and a comparison with three-dimensional (3D) body shape analysis, which offers a promising venue for capturing these complexities. We found that traditional indices systematically misclassified obesity-related risk in certain ancestry groups, with WHR and WHtR showing ancestry-specific biases. In contrast, 3D body shape promises to capture nuanced variations in fat distribution and reduced ancestry-related misclassification. By leveraging techniques based on advanced geometric morphometry and image and data processing, we can better characterize the interaction between genetic ancestry and body composition, ultimately improving the accuracy of obesity diagnosis and stratification in Latin American populations. These results highlight the need for ancestry-aware obesity diagnostics and demonstrate that integrating advanced 3D morphometric techniques can improve risk assessment and guide precision public health strategies in Latin America and beyond. We demonstrate that incorporating 3D body shape data alongside genetic ancestry data improves the accuracy of obesity risk stratification in Latin American populations. Our proposed methods could be adapted, expanded and applied to other populations.- 24 October 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-16
Open AccessGenetic ancestry influences body shape and obesity risk in Latin American populations
Obesity is not simply a matter of excess weight. It also involves changes in structure and proportion in body morphology that can vary between populations and within individuals as they develop and age. Anthropometric measurements and their derived indices are widely used to study obesity. However, they present limitations to capture variations of fat distribution in the human body within a given population, and among different populations. Particularly, currently a problem in epidemiology is that cut-off points and health risk classifications based on anthropometric measures such as BMI, WHR or WHtR may not be equally valid for all population groups, especially when there are differences in genetic ancestry. Using data from \(\sim 7,000\) Latin American adults, we evaluated the accuracy of traditional indices across gradients of Native American, European, and African ancestry, and a comparison with three-dimensional (3D) body shape analysis, which offers a promising venue for capturing these complexities. We found that traditional indices systematically misclassified obesity-related risk in certain ancestry groups, with WHR and WHtR showing ancestry-specific biases. In contrast, 3D body shape promises to capture nuanced variations in fat distribution and reduced ancestry-related misclassification. By leveraging techniques based on advanced geometric morphometry and image and data processing, we can better characterize the interaction between genetic ancestry and body composition, ultimately improving the accuracy of obesity diagnosis and stratification in Latin American populations. These results highlight the need for ancestry-aware obesity diagnostics and demonstrate that integrating advanced 3D morphometric techniques can improve risk assessment and guide precision public health strategies in Latin America and beyond. We demonstrate that incorporating 3D body shape data alongside genetic ancestry data improves the accuracy of obesity risk stratification in Latin American populations. Our proposed methods could be adapted, expanded and applied to other populations. - Open Access
- 18 December 2024
- Published by Springer Nature in Scientific Data
- Vol. 11 (1) , 1-8
3DPatBody: 3D dataset of human bodies of a patagonian population and their anthropometric measurements
The study of human body shape using classical anthropometric techniques is often problematic due to several error sources. Instead, 3D models and representations provide more accurate registrations, which are stable across acquisitions, and enable more precise, systematic, and fast measuring capabilities. Thus, the same person can be scanned several times and precise differential measurements can be established in an accurate manner. Here we present 3DPatBody, a dataset including 3D body scans, with their corresponding 3D point clouds and anthropometric measurements, from a sample of a Patagonian population (female=211, male=87, other=1). The sample is of scientific interest since it is representative of a phenotype characterized by both its biomedical meaning as a descriptor of overweight and obesity, and its population-specific nature related to ancestry and/or local environmental factors. The acquired 3D models were used to compare shape variables against classical anthropometric data. The shape indicators proved to be accurate predictors of classical indices, also adding geometric characteristics that reflect more properly the shape of the body under study.
- 18 December 2024
- Published by Springer Nature in Scientific Data
- Vol. 11 (1) , 1-8
Open Access3DPatBody: 3D dataset of human bodies of a patagonian population and their anthropometric measurements
The study of human body shape using classical anthropometric techniques is often problematic due to several error sources. Instead, 3D models and representations provide more accurate registrations, which are stable across acquisitions, and enable more precise, systematic, and fast measuring capabilities. Thus, the same person can be scanned several times and precise differential measurements can be established in an accurate manner. Here we present 3DPatBody, a dataset including 3D body scans, with their corresponding 3D point clouds and anthropometric measurements, from a sample of a Patagonian population (female=211, male=87, other=1). The sample is of scientific interest since it is representative of a phenotype characterized by both its biomedical meaning as a descriptor of overweight and obesity, and its population-specific nature related to ancestry and/or local environmental factors. The acquired 3D models were used to compare shape variables against classical anthropometric data. The shape indicators proved to be accurate predictors of classical indices, also adding geometric characteristics that reflect more properly the shape of the body under study.
- Open Access
- 6 November 2024
- Published by Instituto Nacional de Salud (Colombia) in Biomédica
- Vol. 44 (4) , 510-523
Relación entre consumo moderado de alcohol, polimorfismos genéticos y peso corporal en una muestra poblacional de Puerto Madryn, Argentina
Introducción. La relación entre la obesidad y el consumo de alcohol es un tema de gran interés para la salud pública. Las bebidas alcohólicas aportan calorías adicionales a la dieta, lo que podría ser un factor relevante en el riesgo de sobrepeso. Sin embargo, su asociación con la ganancia de peso es controversial y está influenciada por múltiples factores. Objetivo. Analizar la relación entre la ingestión moderada de alcohol y el índice de masa corporal, y las variables que pueden influir en dicha relación.Materiales y métodos. La muestra estuvo constituida por 155 personas de Puerto Madryn (Argentina). Cada participante contestó un cuestionario sobre salud, estilo de vida, factores demográficos y socioeconómicos. Se tomaron medidas antropométricas y se tipificaron los polimorfismos de 18 genes relacionados con el metabolismo del alcohol. Resultados. Se encontró que el consumo moderado de alcohol está asociado con un índice de masa corporal más bajo, particularmente en el sexo femenino. Un aumento de 14 gramos de alcohol estuvo asociado con un riesgo de 0,68 para la obesidad y 0,71 para el sobrepeso. La variante T del marcador rs4646543 (ALDH1A1), un gen involucrado en el metabolismo del alcohol y en la adipogénesis, se asoció con una mayor frecuencia de consumo de bebidas alcohólicas.Conclusión. Los hallazgos del presente trabajo sugieren que el consumo moderado de alcohol no contribuye significativamente al peso corporal en la muestra estudiada. Además, la asociación con ciertas variantes genéticas, como las del gen ALDH1A1, podría explicar biológicamente la relación inversa observada entre el peso y el consumo de alcohol.
- 6 November 2024
- Published by Instituto Nacional de Salud (Colombia) in Biomédica
- Vol. 44 (4) , 510-523
Open AccessRelación entre consumo moderado de alcohol, polimorfismos genéticos y peso corporal en una muestra poblacional de Puerto Madryn, Argentina
Introducción. La relación entre la obesidad y el consumo de alcohol es un tema de gran interés para la salud pública. Las bebidas alcohólicas aportan calorías adicionales a la dieta, lo que podría ser un factor relevante en el riesgo de sobrepeso. Sin embargo, su asociación con la ganancia de peso es controversial y está influenciada por múltiples factores. Objetivo. Analizar la relación entre la ingestión moderada de alcohol y el índice de masa corporal, y las variables que pueden influir en dicha relación.Materiales y métodos. La muestra estuvo constituida por 155 personas de Puerto Madryn (Argentina). Cada participante contestó un cuestionario sobre salud, estilo de vida, factores demográficos y socioeconómicos. Se tomaron medidas antropométricas y se tipificaron los polimorfismos de 18 genes relacionados con el metabolismo del alcohol. Resultados. Se encontró que el consumo moderado de alcohol está asociado con un índice de masa corporal más bajo, particularmente en el sexo femenino. Un aumento de 14 gramos de alcohol estuvo asociado con un riesgo de 0,68 para la obesidad y 0,71 para el sobrepeso. La variante T del marcador rs4646543 (ALDH1A1), un gen involucrado en el metabolismo del alcohol y en la adipogénesis, se asoció con una mayor frecuencia de consumo de bebidas alcohólicas.Conclusión. Los hallazgos del presente trabajo sugieren que el consumo moderado de alcohol no contribuye significativamente al peso corporal en la muestra estudiada. Además, la asociación con ciertas variantes genéticas, como las del gen ALDH1A1, podría explicar biológicamente la relación inversa observada entre el peso y el consumo de alcohol.
- Open Access
- 29 December 2023
- Published by Universidad Nacional de La Plata in Revista Argentina de Antropología Biológica
- Vol. 26 (1) , 073
Influencia de variables genéticas y socio-ambientales en el exceso de peso en una muestra de la población de Puerto Madryn, Argentina
En Argentina, el exceso de peso aumentó continuamente desde el 2005 hasta el presente. Los factores con probada relación con el exceso de peso son socio-ambientales y genéticos. Entre estos últimos, el gen FTO es el más estudiado. El objetivo de este trabajo fue analizar la relación entre dos marcadores del gen FTO (rs9939609 y rs11075997) y variables socio-ambientales con el estado nutricional en una muestra de 155 personas de Puerto Madryn, Chubut. Se calcularon los índices de masa corporal (IMC) y de cintura talla (ICT), y se probaron modelos lineales múltiples para conocer la influencia de las variables genéticas y no genéticas sobre dichos índices antropométricos. El nivel educativo y la actividad física estuvieron asociados significativamente y de manera negativa con el estado nutricional. El alelo reportado en la literatura como de riesgo del marcador rs11075997 estuvo significativamente asociado con el aumento de los índices antropométricos bajo un modelo aditivo, en mayor medida con el ICT que con el IMC. Es necesario realizar trabajos a mayor escala, que permitan identificar la frecuencia de los marcadores genéticos de predisposición a obesidad aquí estudiados y su distribución en nuestro país.
- 29 December 2023
- Published by Universidad Nacional de La Plata in Revista Argentina de Antropología Biológica
- Vol. 26 (1) , 073
Open AccessInfluencia de variables genéticas y socio-ambientales en el exceso de peso en una muestra de la población de Puerto Madryn, Argentina
En Argentina, el exceso de peso aumentó continuamente desde el 2005 hasta el presente. Los factores con probada relación con el exceso de peso son socio-ambientales y genéticos. Entre estos últimos, el gen FTO es el más estudiado. El objetivo de este trabajo fue analizar la relación entre dos marcadores del gen FTO (rs9939609 y rs11075997) y variables socio-ambientales con el estado nutricional en una muestra de 155 personas de Puerto Madryn, Chubut. Se calcularon los índices de masa corporal (IMC) y de cintura talla (ICT), y se probaron modelos lineales múltiples para conocer la influencia de las variables genéticas y no genéticas sobre dichos índices antropométricos. El nivel educativo y la actividad física estuvieron asociados significativamente y de manera negativa con el estado nutricional. El alelo reportado en la literatura como de riesgo del marcador rs11075997 estuvo significativamente asociado con el aumento de los índices antropométricos bajo un modelo aditivo, en mayor medida con el ICT que con el IMC. Es necesario realizar trabajos a mayor escala, que permitan identificar la frecuencia de los marcadores genéticos de predisposición a obesidad aquí estudiados y su distribución en nuestro país.
- Open Access
- 1 May 2023
- Published by AIP Publishing in Chaos: An Interdisciplinary Journal of Nonlinear Science
- Vol. 33 (5)
Ordinal pattern transition networks in eye tracking reading signals
Eye tracking is an emerging technology with a wide spectrum of applications, including non-invasive neurocognitive diagnosis. An advantage of the use of eye trackers is in the improved assessment of indirect latent information about several aspects of the subjects’ neurophysiology. The path to uncover and take advantage of the meaning and implications of this information, however, is still in its very early stages. In this work, we apply ordinal patterns transition networks as a means to identify subjects with dyslexia in simple text reading experiments. We registered the tracking signal of the eye movements of several subjects (either normal or with diagnosed dyslexia). The evolution of the left-to-right movement over time was analyzed using ordinal patterns, and the transitions between patterns were analyzed and characterized. The relative frequencies of these transitions were used as feature descriptors, with which a classifier was trained. The classifier is able to distinguish typically developed vs dyslexic subjects with almost 100% accuracy only analyzing the relative frequency of the eye movement transition from one particular permutation pattern (plain left to right) to four other patterns including itself. This characterization helps understand differences in the underlying cognitive behavior of these two groups of subjects and also paves the way to several other potentially fruitful analyses applied to other neurocognitive conditions and tests.
- 1 May 2023
- Published by AIP Publishing in Chaos: An Interdisciplinary Journal of Nonlinear Science
- Vol. 33 (5)
Open AccessOrdinal pattern transition networks in eye tracking reading signals
Eye tracking is an emerging technology with a wide spectrum of applications, including non-invasive neurocognitive diagnosis. An advantage of the use of eye trackers is in the improved assessment of indirect latent information about several aspects of the subjects’ neurophysiology. The path to uncover and take advantage of the meaning and implications of this information, however, is still in its very early stages. In this work, we apply ordinal patterns transition networks as a means to identify subjects with dyslexia in simple text reading experiments. We registered the tracking signal of the eye movements of several subjects (either normal or with diagnosed dyslexia). The evolution of the left-to-right movement over time was analyzed using ordinal patterns, and the transitions between patterns were analyzed and characterized. The relative frequencies of these transitions were used as feature descriptors, with which a classifier was trained. The classifier is able to distinguish typically developed vs dyslexic subjects with almost 100% accuracy only analyzing the relative frequency of the eye movement transition from one particular permutation pattern (plain left to right) to four other patterns including itself. This characterization helps understand differences in the underlying cognitive behavior of these two groups of subjects and also paves the way to several other potentially fruitful analyses applied to other neurocognitive conditions and tests.
- 6 February 2023
- Published by Wiley in American Journal of Biological Anthropology
- Vol. 181 (1) , 85-95
Genetic and self‐perceived ancestries in Argentina: Beyond the three‐hybrid model
Objectives: The increased availability of genome‐wide data allows capturing the fine genetic structure of present days populations. Here we analyze the genetic ancestry at a fine scale of an Argentinean Patagonia population to understand the origins beyond the three‐hybrid model, and to compare these results with volunteers' self‐perceived ancestry in a broad context encompassed by historical and familiar information.Materials and Methods: We compare high‐throughput genotyping data for 92 individuals that we generated to data sets from the literature by applying fully haplotype‐based methods to examine patterns of human population substructure. The volunteers filled out a semi‐structured questionnaire, including questions about their history, ancestors, and self‐perceived ancestry. Finally, we used non‐parametric tests in order to compare genomic ancestry against self‐perception.Results: Genetic ancestry from Iberian populations accounted for 0.176 (Spain and Basque origins), while the component associated with Italian populations accounted for 0.140. We observed a 0.169 Native American genetic ancestry. Participants significantly over‐ and under‐ self‐perceived Native American and European origins, respectively. Components of origins from North Africa to Central South Asia accounted for 0.225 of the genetic ancestry in the sample, with significantly higher proportions for people that mentioned such origins in their genealogical history.Discussion: We captured the fine‐genetic architecture of a Puerto Madryn population sample in Chubut province, showing that self‐perceived ancestry remains a poor proxy for genetic ancestry. The presence of North Africa to Central South Asia components and its correlate with self‐perception of these origins justifies its inclusion in future miscegenation studies in Argentina.
- 6 February 2023
- Published by Wiley in American Journal of Biological Anthropology
- Vol. 181 (1) , 85-95
Genetic and self‐perceived ancestries in Argentina: Beyond the three‐hybrid model
Objectives: The increased availability of genome‐wide data allows capturing the fine genetic structure of present days populations. Here we analyze the genetic ancestry at a fine scale of an Argentinean Patagonia population to understand the origins beyond the three‐hybrid model, and to compare these results with volunteers' self‐perceived ancestry in a broad context encompassed by historical and familiar information.Materials and Methods: We compare high‐throughput genotyping data for 92 individuals that we generated to data sets from the literature by applying fully haplotype‐based methods to examine patterns of human population substructure. The volunteers filled out a semi‐structured questionnaire, including questions about their history, ancestors, and self‐perceived ancestry. Finally, we used non‐parametric tests in order to compare genomic ancestry against self‐perception.Results: Genetic ancestry from Iberian populations accounted for 0.176 (Spain and Basque origins), while the component associated with Italian populations accounted for 0.140. We observed a 0.169 Native American genetic ancestry. Participants significantly over‐ and under‐ self‐perceived Native American and European origins, respectively. Components of origins from North Africa to Central South Asia accounted for 0.225 of the genetic ancestry in the sample, with significantly higher proportions for people that mentioned such origins in their genealogical history.Discussion: We captured the fine‐genetic architecture of a Puerto Madryn population sample in Chubut province, showing that self‐perceived ancestry remains a poor proxy for genetic ancestry. The presence of North Africa to Central South Asia components and its correlate with self‐perception of these origins justifies its inclusion in future miscegenation studies in Argentina.
- Open Access
- 11 September 2020
- Published by MDPI AG in Journal of Imaging
- Vol. 6 (9) , 94
body2vec: 3D Point Cloud Reconstruction for Precise Anthropometry with Handheld Devices
Current point cloud extraction methods based on photogrammetry generate large amounts of spurious detections that hamper useful 3D mesh reconstructions or, even worse, the possibility of adequate measurements. Moreover, noise removal methods for point clouds are complex, slow and incapable to cope with semantic noise. In this work, we present body2vec, a model-based body segmentation tool that uses a specifically trained Neural Network architecture. Body2vec is capable to perform human body point cloud reconstruction from videos taken on hand-held devices (smartphones or tablets), achieving high quality anthropometric measurements. The main contribution of the proposed workflow is to perform a background removal step, thus avoiding the spurious points generation that is usual in photogrammetric reconstruction. A group of 60 persons were taped with a smartphone, and the corresponding point clouds were obtained automatically with standard photogrammetric methods. We used as a 3D silver standard the clean meshes obtained at the same time with LiDAR sensors post-processed and noise-filtered by expert anthropological biologists. Finally, we used as gold standard anthropometric measurements of the waist and hip of the same people, taken by expert anthropometrists. Applying our method to the raw videos significantly enhanced the quality of the results of the point cloud as compared with the LiDAR-based mesh, and of the anthropometric measurements as compared with the actual hip and waist perimeter measured by the anthropometrists. In both contexts, the resulting quality of body2vec is equivalent to the LiDAR reconstruction.
- 11 September 2020
- Published by MDPI AG in Journal of Imaging
- Vol. 6 (9) , 94
Open Accessbody2vec: 3D Point Cloud Reconstruction for Precise Anthropometry with Handheld Devices
Current point cloud extraction methods based on photogrammetry generate large amounts of spurious detections that hamper useful 3D mesh reconstructions or, even worse, the possibility of adequate measurements. Moreover, noise removal methods for point clouds are complex, slow and incapable to cope with semantic noise. In this work, we present body2vec, a model-based body segmentation tool that uses a specifically trained Neural Network architecture. Body2vec is capable to perform human body point cloud reconstruction from videos taken on hand-held devices (smartphones or tablets), achieving high quality anthropometric measurements. The main contribution of the proposed workflow is to perform a background removal step, thus avoiding the spurious points generation that is usual in photogrammetric reconstruction. A group of 60 persons were taped with a smartphone, and the corresponding point clouds were obtained automatically with standard photogrammetric methods. We used as a 3D silver standard the clean meshes obtained at the same time with LiDAR sensors post-processed and noise-filtered by expert anthropological biologists. Finally, we used as gold standard anthropometric measurements of the waist and hip of the same people, taken by expert anthropometrists. Applying our method to the raw videos significantly enhanced the quality of the results of the point cloud as compared with the LiDAR-based mesh, and of the anthropometric measurements as compared with the actual hip and waist perimeter measured by the anthropometrists. In both contexts, the resulting quality of body2vec is equivalent to the LiDAR reconstruction.