Publicado:
2026-07-01Número:
Vol. 20 Núm. 1 (2026)Sección:
Visión InvestigadoraFringe pattern images for disruptive structured data classification
Imágenes de franjas para clasificación disruptiva de datos estructurados
Palabras clave:
Data science, structured data, visual representation, fringe patterns, deep learning (en).Palabras clave:
Ciencia de datos, datos estructurados, representación visual, patrones de franjas, aprendizaje profundo (es).Descargas
Resumen (en)
Conventional models often fall short in supporting decision-making from structured tabular data, as they struggle to capture complex feature interactions and nonlinear dependencies. In contrast, convolutional neural networks (CNNs), highly effective in computer vision, remain underutilized in structured data analysis. This work introduces a disruptive data science approach that bridges this gap by transforming tabular records into fringe pattern images through Gaussian surface modeling. These images encode feature intensity, spatial frequency, and phase-based texture, enabling CNNs to exploit new visual cues beyond raw numerical inputs. The method is systematically evaluated across architectures such as ResNet, DenseNet, and GoogleNet using the PIMA Indians Diabetes dataset as a benchmark. Experimental results show that the proposed fringe-based visual representation consistently outperforms baseline classifiers, achieving higher accuracy and richer feature discrimination. By reframing structured data as visual information, this approach demonstrates how disruptive applications of data science can enhance healthcare, finance, and industrial analytics, fostering models that are both more accurate and more interpretable.
Resumen (es)
Los modelos convencionales de clasificación suelen ser insuficientes para apoyar la toma de decisiones con datos estructurados en formato tabular, ya que presentan limitaciones al capturar interacciones complejas y dependencias no lineales. En contraste, las redes neuronales convolucionales (CNN), altamente efectivas en visión por computador, permanecen subutilizadas en el análisis de datos estructurados. Este trabajo introduce un enfoque disruptivo en ciencia de datos que transforma registros tabulares en imágenes de patrones de franjas mediante modelado de superficies gaussianas. Estas imágenes codifican intensidad de atributos, frecuencia espacial y textura basada en fase, lo que permite a las CNN aprovechar nuevas señales visuales más allá de los valores numéricos originales. El método se evalúa en arquitecturas como ResNet, DenseNet y GoogleNet utilizando como referencia el conjunto de datos PIMA Indians Diabetes. Los resultados experimentales muestran que la representación visual propuesta supera de forma consistente a los clasificadores convencionales, logrando mayor precisión y discriminación de características. Al replantear los datos estructurados como información visual, este enfoque evidencia cómo las aplicaciones disruptivas de la ciencia de datos pueden potenciar áreas como la salud, las finanzas y la analítica industrial, fomentando modelos más precisos e interpretables.
Referencias
[1] G. Saranya and S. D. Pande, “Enhancing diabetes prediction with data preprocessing and various machine learning algorithms,” EAI Endorsed Transactions on Internet of Things, vol. 10, 2024.
[2] N. Sneha and T. Gangil, “Analysis of diabetes mellitus for early prediction using optimal features selection,” Journal of Big Data, vol. 6, no. 1, pp. 1–19, 2019.
[3] J. R. Quinlan, “Induction of decision trees,” Machine Learning, vol. 1, no. 1, pp. 81–106, 1986.
[4] R. G. Brereton and G. R. Lloyd, “Support vector machines for classification and regression,” The Analyst, vol. 135, no. 2, pp. 230–267, 2010.
[5] M. F. Aslan and K. Sabanci, “A novel proposal for deep learning-based diabetes prediction: converting clinical data to image data,” Diagnostics, vol. 13, no. 4, p. 796, 2023.
[6] E. M. F. Donayre, M. L. G. Guevara, R. H. Flores, D. J. C. Moscoso, A. A. S. Macalupu, and M. E. H. Flores, “Machine learning en la Industria 4.0: Análisis de su relevancia y aplicaciones,” Ñawparisun - Revista de Investigación Científica de Ingenierías, vol. 4, no. 3, pp. 93–98, 2023.
[7] Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner, “Gradient-based learning applied to document recognition,” Proceedings of the IEEE, vol. 86, no. 11, pp. 2278–2324, 1998.
[8] A. Krizhevsky, I. Sutskever, and G. E. Hinton, “ImageNet classification with deep convolutional neural networks,” Advances in Neural Information Processing Systems, vol. 25, 2012.
[9] N. Hatami, Y. Gavet, and J. Debayle, “Classification of time-series images using deep convolutional neural networks,” in Tenth International Conference on Machine Vision (ICMV 2017), vol. 10696, SPIE, 2018, pp. 242–249.
[10] N. R. B. S. Loka, M. Kavitha, and T. Kurita, “Hilbert vector convolutional neural network: 2D neural network on 1D data,” in Proc. Int. Conf. Artificial Neural Networks, 2019, pp. 458–470.
[11] J. Yan, Y. Huang, S. Yuan, Y. Lu, and Z. Yu, “Open-circuit fault analysis and recognition in three-level inverters based on recurrence plot and convolution neural network,” Int. Trans. Electr. Energy Syst., vol. 2023, no. 1, p. 4755960, 2023.
[12] A. Mousa, W. Mustafa, R. B. Marqas, and S. H. M. Mohammed, “A comparative study of diabetes detection using the Pima Indian diabetes database,” Journal of Duhok University, vol. 26, no. 2, pp. 277–288, 2023.
[13] P. Ma, H. Zhang, W. Fan, C. Wang, G. Wen, and X. Zhang, “A novel bearing fault diagnosis method based on 2D image representation and transfer learning-convolutional neural network,” Measurement Science and Technology, vol. 30, no. 5, p. 055402, 2019.
[14] J.-A. Lee and K.-C. Kwak, “Personal identification using an ensemble approach of 1D-LSTM and 2D-CNN with electrocardiogram signals,” Applied Sciences, vol. 12, no. 5, p. 2692, 2022.
[15] M. O. Gutiérrez, L. M. H. Ramírez, M. A. S. Verduzco, M. P. Cortés, and J. C. I. Torres, “Proyección de luz estructurada para la reconstrucción de superficies en 3D,” Ciencia Nicolaita, no. 66, pp. 84–96, 2015.
[16] LI, Beiwen. High-speed 3d optical sensing for manufacturing research and industrial sensing applications. Transactions on Energy Systems and Engineering Applications, 2022, vol. 3, no 2, p. 1-12.
[17] Briñez-De-Leon J-C, Rico-Garcia M, Restrepo-Martínez A. Isochromatic-Art: A Computational Dataset for Digital Photoelasticity Studies. Data. 2022; 7(11):151. https://doi.org/10.3390/data7110151.
[18] K. Simonyan, “Very deep convolutional networks for large-scale image recognition,” arXiv preprint arXiv:1409.1556, 2014.
[19] C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V. Vanhoucke, and A. Rabinovich, “Going deeper with convolutions,” in Proc. IEEE Conf. Comput. Vision Pattern Recognit., 2015, pp. 1–9.
[20] D. Renza and D. M. Ballesteros, Fundamentos de visión por computador utilizando aprendizaje profundo, Cali, Colombia: Editorial Redipe, 2023.
[21] J. Brownlee, Deep learning for computer vision: image classification, object detection, and face recognition in Python, Machine Learning Mastery, 2019.
[22] C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V. Vanhoucke, and A. Rabinovich, “Going deeper with convolutions,” in Proc. IEEE Conf. Comput. Vision Pattern Recognit., 2015, pp. 1–9.
[23] G. Huang, Z. Liu, L. Van Der Maaten, and K. Q. Weinberger, “Densely connected convolutional networks,” in Proc. IEEE Conf. Comput. Vision Pattern Recognit., 2017, pp. 4700–4708.
Cómo citar
APA
ACM
ACS
ABNT
Chicago
Harvard
IEEE
MLA
Turabian
Vancouver
Descargar cita
Visitas
Descargas
Licencia
Esta obra está bajo una licencia internacional Creative Commons Atribución-NoComercial 4.0.
.png)
atribución- no comercial 4.0 International
.jpg)
