Image Feature Extraction Using Matrix Calculus Techniques: Applications in Texture Analysis and Pattern Recognition

Ccama Alejo R.; Marlon Mollinedo Chura R.; Ticona Huayhua R.; Vilca Callata L. · Proceedings of the LACCEI international Multi-conference for Engineering, Education and Technology (2026) · DOI: 10.18687/laccei2026.1.1.489

Resumen

Feature extraction is a fundamental step in digital image processing, as it allows the transformation of high-dimensional visual information into compact and discriminative representations for pattern analysis. This paper proposes a feature extraction model based on matrix calculus and linear algebra techniques applied to digital images. The images are represented as numerical matrices, enabling the use of mathematical tools such as the Discrete Fourier Transform (DFT), the Discrete Cosine Transform (DCT), Singular Value Decomposition (SVD), Principal Component Analysis (PCA), and Globally Connected Matrices (GLCM). These methodologies allow the analysis of spectral, structural, and textural properties of the images, facilitating dimensionality reduction, redundancy elimination, and the identification of relevant patterns. The proposed methodology comprises stages of matrix representation, preprocessing, mathematical transformation, and extraction of numerical descriptors oriented toward texture analysis and pattern recognition. The results demonstrate that matrix computation techniques provide efficient representations of visual information, standing out for their mathematical interpretability, computational efficiency, and applicability to various computer vision and digital image processing problems.

Instituciones: Universidad Nacional del Altiplano