Análisis de imágenes en platos de comida típicos de Antioquía: un enfoque de visión por computadora para el conteo de carbohidratos en personas con diabetes mellitus.

dc.contributor.advisorOspina Arango, Juan David
dc.contributor.advisorBolaños Martínez, Freddy
dc.contributor.advisorLema Pérez, Laura
dc.contributor.authorJaramillo Tamayo, Andrés Felipe
dc.date.accessioned2025-10-07T21:05:02Z
dc.date.available2025-10-07T21:05:02Z
dc.date.issued2025-08-18
dc.descriptionIlustraciones
dc.description.abstractEsta investigación tuvo como objetivo implementar un sistema computacional basado en visión por computador para cuantificar gramos de carbohidratos en platos típicos de Antioquia, con el fin de apoyar el control en personas con diabetes tipo 1. Se construyó un banco de datos compuesto por 1.328 imágenes anotadas manualmente, representando 41 tipos de carbohidratos. Se entrenaron y evaluaron tres modelos de segmentación semántica multiclase, destacando la aplicación de transfer learning de UperNet con backbone ConvNeXt, el cual alcanzó un mIoU de validación de 0.758 y una precisión del 95.3 %. Para la conversión de área segmentada a gramos, se integró la metodología Look Up Table (LUT), que relaciona el porcentaje de cobertura del alimento respecto al plato con su masa estimada, usando el ATLAS como referencia. Esta aproximación presentó errores absolutos promedio entre el 35 % y el 45 %, validados mediante pesajes gravimétricos en tres platos reales. El sistema demostró mayor precisión en clases con alta representación visual y evidencia un alto potencial de aplicabilidad en contextos clínicos y educativos. Se recomienda ampliar la base de datos, mejorar la representación por clase y automatizar el proceso de etiquetado para aumentar la escalabilidad del sistema. (Texto tomado de la fuente)spa
dc.description.abstractThis research aimed to implement a computer vision-based system to quantify grams of carbohydrates in traditional dishes from Antioquia, with the goal of supporting control in individuals with type 1 diabetes. A dataset of 1,328 manually annotated images was constructed, representing 41 types of carbohydrates. Three multiclass semantic segmentation models were trained and evaluated, highlighting the application of transfer learning through UperNet with a ConvNeXt backbone, which achieved a validation mIoU of 0.758 and an accuracy of 95.3%. For the conversion of segmented area to grams, the Look Up Table (LUT) methodology was integrated, which relates the percentage of food coverage on the plate to its estimated mass, using the ATLAS as a reference. This approach yielded average absolute errors ranging between 35% and 45%, validated through gravimetric measurements on three real plates. The system demonstrated higher precision for classes with strong visual representation and shows high potential for clinical and educational applications. It is recommended to expand the dataset, improve class representation, and automate the annotation process to enhance system scalability.eng
dc.description.curricularareaIngeniería de Sistemas e Informática
dc.description.degreelevelMaestría
dc.description.degreenameMaestría en Ingeniería - Analítica
dc.description.researchareaIngeniería biomédica, Ingeniería analítica, Ingeniería química, Salud
dc.description.technicalinfohttps://github.com/andresjaramillotamayo/m-health-carbs-cuantification Github donde vive todo el modelospa
dc.format.extent1 recurso en líne (86 páginas)
dc.format.mimetypeapplication/pdf
dc.identifier.instnameUniversidad Nacional de Colombiaspa
dc.identifier.reponameRepositorio Institucional Universidad Nacional de Colombiaspa
dc.identifier.repourlhttps://repositorio.unal.edu.co/spa
dc.identifier.urihttps://repositorio.unal.edu.co/handle/unal/89017
dc.language.isospa
dc.publisherUniversidad Nacional de Colombia
dc.publisher.branchUniversidad Nacional de Colombia - Sede Medellín
dc.publisher.facultyFacultad de Minas
dc.publisher.placeMedellín, Colombia
dc.publisher.programMedellín - Minas - Maestría en Ingeniería - Analítica
dc.relation.referencesAlfonsi, J. E., Choi, E. E. Y ., Arshad, T., Sammott, S. A. S., Pais, V ., Nguyen, C., Maguire, B. R., Stinson, J. N., & Palmert, M. R. (2020). Carbohydrate counting app using image recognition for youth with Type 1 diabetes: Pilot randomized control trial. JMIR MHealth and UHealth, 8(10). https://doi.org/10.2196/22074
dc.relation.referencesAnthimopoulos, M., Dehais, J., Shevchik, S., Ransford, B. H., Duke, D., Diem, P., & Mougiakakou, S. (2015). Computer vision-based carbohydrate estimation for type 1 patients with diabetes using smartphones. Journal of Diabetes Science and Technology, 9(3), 507–515. https://doi.org/10.1177/1932296815580159
dc.relation.referencesAnthimopoulos, M., Dehais, J., Shevchik, S., Ransford, B. H., Duke, D., Diem, P., & Mougiakakou, S. (2015). Computer vision-based carbohydrate estimation for type 1 patients with diabetes using smartphones. Journal of Diabetes Science and Technology, 9(3), 507–515. https://doi.org/10.1177/1932296815580159
dc.relation.referencesAnthimopoulos, M., Dehais, J., Shevchik, S., Ransford, B. H., Duke, D., Diem, P., & Mougiakakou, S. (2015). Computer vision-based carbohydrate estimation for type 1 patients with diabetes using smartphones. Journal of Diabetes Science and Technology, 9(3), 507–515. https://doi.org/10.1177/1932296815580159
dc.relation.referencesCaggia, A. (2021, December 27). We Tried the SNAQ Diabetes App. Diabetes Daily. https://www.diabetesdaily.com/blog/we-tried-the-snaq-diabetes-app-695058/
dc.relation.referencesCanny, J. (1986). A Computational Approach to Edge Detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 8(6), 679–698. https://doi.org/10.1109/TPAMI.1986.4767851
dc.relation.referencesChen, L.-C., Zhu, Y ., Papandreou, G., Schroff, F., & Adam, H. (2018). Encoder-decoder with atrous separable convolution for semantic image segmentation (arXiv preprint arXiv:1807.10221). arXiv. https://arxiv.org/abs/1807.10221
dc.relation.referencesCortes, C., & Vapnik, V . (1995). Support-vector networks. Machine learning, 20(3), 273–297. https://doi.org/10.1007/BF00994018
dc.relation.referencesDepaoli, R., Fernández, L. A., & Díaz, D. (s.f.). Optimización de la ecualización del histograma en el procesamiento de imágenes digitales. Universidad Nacional de La Matanza. Recuperado de https://sedici.unlp.edu.ar/bitstream/handle/10915/21082/51.pdf?sequence=1
dc.relation.referencesElngar, A. A. (2021). New calculation of calorie content and determining nutritional level from day-to-day intake of food using Image Processing. https://ssrn.com/abstract=4226672
dc.relation.referencesGeeksforGeeks. (n.d.). Difference Between RGB, CMYK, HSV , and YIQ Color Models. GeeksforGeeks. Recuperado de https://www.geeksforgeeks.org/difference-between-rgb-cmyk-hsv-and-yiq-color-models/
dc.relation.referencesGrupo de Investigación Alimentación y Nutrición Humana (GIANH), Escuela de Nutrición y Dietética, Universidad de Antioquia. (2023). Lista de Intercambios (5.ª ed.). Universidad de Antioquia.
dc.relation.referencesHuang, J., Ding, H., McBride, S., Ireland, D., & Karunanithi, M. (2015). Use of Smartphones to Estimate Carbohydrates in Foods for Diabetes Management. Studies in Health Technology and Informatics, 214, 121–127. https://doi.org/10.3233/978-1-61499-558-6-121
dc.relation.referencesInstitute of Electrical and Electronics Engineers. (n.d.). 2018 22nd International Computer Science and Engineering Conference (ICSEC).
dc.relation.referencesIstepanian, R. S. H., & Al-Anzi, T. (2018). m-Health 2.0: New perspectives on mobile health, machine learning and big data analytics. Methods, 151, 34–40. https://doi.org/10.1016/j.ymeth.2018.05.015.
dc.relation.referencesJiang, L., Qiu, B., Liu, X., Huang, C., & Lin, K. (2020). DeepFood: Food Image Analysis and Dietary Assessment via Deep Model. IEEE Access, 8, 47477–47489. https://doi.org/10.1109/ACCESS.2020.2973625
dc.relation.referencesKhan, Z. F., & Alotaibi, S. R. (2020). Applications of Artificial Intelligence and Big Data Analytics in m-Health: A Healthcare System Perspective. In Journal of Healthcare Engineering (V ol. 2020). Hindawi Limited. https://doi.org/10.1155/2020/8894694
dc.relation.referencesKirillov, A., Mintun, E., Ravi, N., Mao, H., Rolland, C., Gustafson, L., Xiao, T., Whitehead, S., Berg, A. C., Lo, W‑Y ., Dollár, P., & Girshick, R. (2023). Segment Anything (arXiv preprint arXiv:2304.02643). arXiv. https://doi.org/10.48550/arXiv.2304.02643
dc.relation.referencesLightDestory. (n.d.). upernetconvnext‑finetuned‑segments‑food‑oct‑14 [Repositorio de modelo de aprendizaje automático]. Hugging Face. https://huggingface.co/LightDestory/upernetconvnext-finetuned-segments-food-oct-14
dc.relation.referencesLiu, Y . C., Onthoni, D. D., Mohapatra, S., Irianti, D., & Sahoo, P. K. (2022). Deep- Learning-Assisted Multi-Dish Food Recognition Application for Dietary Intake Reporting. Electronics (Switzerland), 11(10). https://doi.org/10.3390/electronics11101626
dc.relation.referencesLiu, Z., Mao, H., Wu, C. Y ., Feichtenhofer, C., Darrell, T., & Xie, S. (2022). A ConvNet for the 2020s. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR),11976–11986. https://openaccess.thecvf.com/content/CVPR2022/html/Liu_A_ConvNet_for_the_2020s_CVPR_ 2022_paper.html
dc.relation.referencesNosotros - Kutai. (2022, September 14). Kutai - https://kutai.co/nosotros/
dc.relation.referencesMacQueen, J. (1967). Some Methods for Classification and Analysis of Multivariate Observations. Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, 281–297
dc.relation.referencesMamani Urrutia, V . A., Durán Galdo, R. F., & Salvatierra Ruiz, R. V . (2022). Atlas fotográfico de alimentos peruanos para evaluar el consumo alimentario de niños de 6 a 12 meses de edad: Una primera aproximación desde la investigación (1.ª ed. digital). Universidad Científica del Sur. https://fondoeditorial.cientifica.edu.pe/
dc.relation.referencesMerchant, K., & Pande, Y . (2019). ConvFood: A CNN-Based Food Recognition Mobile Application for Obese and Diabetic Patients. Advances in Intelligent Systems and Computing, 882, 493–502. https://doi.org/10.1007/978-981-13-5953-8_41
dc.relation.referencesMicrosoft. (n.d.). beit-base-patch16-224 [Repositorio de modelo de aprendizaje automático]. Hugging Face. https://huggingface.co/microsoft/beit-base-patch16-224
dc.relation.referencesMinisterio de Salud y Protección Social. (2019). En el Día Mundial de la Diabetes: MinSalud promueve prácticas de vida saludable. Minsalud.gov.co. https://www.minsalud.gov.co/Paginas/En-el-Dia-Mundial-de-la-Diabetes-MinSalud-promueve-pra ticas-de-vida-saludable.aspx.
dc.relation.referencesMinh V o, L., Wang, W., Liu, J., & Li, X. (2021). V olume estimation from single-view images using deep learning and computer vision techniques. International Conference on Computer Vision (ICCV). Recuperado de https://sites.google.com/view/cvpr-metafood-2024/overview?authuser=0
dc.relation.referencesOtsu, N. (1979). A threshold selection method from gray-level histograms. IEEE Transactions on Systems, Man, and Cybernetics, 9(1), 62–66. https://doi.org/10.1109/TSMC.1979.4310076
dc.relation.referencesPerez, S. (2015, February 4). Under Armour Snatches Up Health And Fitness Trackers Endomondo And MyFitnessPal | TechCrunch. TechCrunch. https://techcrunch.com/2015/02/04/athletic-apparel-company-under-armour-snatches-up-health-an d-fitness-trackers-endomondo-and-myfitnesspal/
dc.relation.referencesRhyner, D., Loher, H., Dehais, J., Anthimopoulos, M., Shevchik, S., Botwey, R. H., Duke, D., Stettler, C., Diem, P., & Mougiakakou, S. (2016). Carbohydrate estimation by a mobile phone-based system versus self-estimations of individuals with type 1 diabetes mellitus: A comparative study. Journal of Medical Internet Research, 18(5). https://doi.org/10.2196/jmir.5567
dc.relation.referencesSaeedi, P., Petersohn, I., Salpea, P., Malanda, B., Karuranga, S., Unwin, N., Colagiuri, S., Guariguata, L., Motala, A. A., Ogurtsova, K., Shaw, J. E., Bright, D., & Williams, R. (2019a). Global and regional diabetes prevalence estimates for 2019 and projections for2030 and 2045: Results from the International Diabetes Federation Diabetes Atlas, 9th edition. Diabetes Research and Clinical Practice, 157. https://doi.org/10.1016/j.diabres.2019.107843
dc.relation.referencesSaavedra Campos, R. (2021). Aplicación de métodos de procesado digital de imagen para la detección de objetos en entornos industriales (Trabajo de fin de grado). Universidade da Coruña. Recuperado de https://ruc.udc.es/dspace/bitstream/handle/2183/31863/SaavedraCampos_Raul_TFG_2021.pdf?se quence=3
dc.relation.referencesSite, A., Nurmi, J., & Lohan, E. S. (2021). Systematic review on machine-learning algorithms used in wearable-based eHealth data analysis. IEEE Access, 9, 112221–112235. https://doi.org/10.1109/ACCESS.2021.3103268
dc.relation.referencesWang, L., & Alexander, C. A. (2020). Big data analytics in medical engineering and healthcare: methods, advances and challenges. In Journal of Medical Engineering and Technology (V ol. 44, Issue 6, pp. 267––283). Taylor and Francis Ltd. https://doi.org/10.1080/03091902.2020.1769758
dc.rights.accessrightsinfo:eu-repo/semantics/openAccess
dc.rights.licenseAtribución-NoComercial 4.0 Internacional
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/
dc.subject.ddc640 - Gestión del hogar y vida familiar::641 - Alimentos y bebidas
dc.subject.proposalSegmentación semánticaspa
dc.subject.proposalVisión por computadorspa
dc.subject.proposalInteligencia artificialspa
dc.subject.proposalM-healthspa
dc.subject.proposalAprendizaje automáticospa
dc.subject.proposalEstimación de carbohidratosspa
dc.subject.proposalDiabetes tipo 1spa
dc.subject.proposalSemantic segmentationeng
dc.subject.proposalComputer visioneng
dc.subject.proposalArtificial intelligenceeng
dc.subject.proposalM-healtheng
dc.subject.proposalMachine learningeng
dc.subject.proposalCarbohydrate estimationeng
dc.subject.proposalType 1 diabeteseng
dc.subject.wikidataSegmentación
dc.subject.wikidataInteligencia artificial
dc.subject.wikidataAprendizaje automático
dc.subject.wikidataCarbohidratos
dc.titleAnálisis de imágenes en platos de comida típicos de Antioquía: un enfoque de visión por computadora para el conteo de carbohidratos en personas con diabetes mellitus.spa
dc.title.translatedImage analysis of typical Antioquia dishes: a computer vision approach for carbohydrate counting in people with diabetes mellitus.eng
dc.typeTrabajo de grado - Maestría
dc.type.coarhttp://purl.org/coar/resource_type/c_bdcc
dc.type.coarversionhttp://purl.org/coar/version/c_ab4af688f83e57aa
dc.type.contentText
dc.type.contentText
dc.type.driverinfo:eu-repo/semantics/masterThesis
dc.type.redcolhttp://purl.org/redcol/resource_type/TM
dc.type.versioninfo:eu-repo/semantics/acceptedVersion
dcterms.audience.professionaldevelopmentInvestigadores
oaire.accessrightshttp://purl.org/coar/access_right/c_abf2

Archivos

Bloque original

Mostrando 1 - 1 de 1
Cargando...
Miniatura
Nombre:
TdeG_MA.ing.Analitica_AndresJaramillo_m_health_1037663920.pdf
Tamaño:
15.82 MB
Formato:
Adobe Portable Document Format

Bloque de licencias

Mostrando 1 - 1 de 1
Cargando...
Miniatura
Nombre:
license.txt
Tamaño:
5.74 KB
Formato:
Item-specific license agreed upon to submission
Descripción: