Department of Mathematical Sciences¹
University of Texas at El Paso
El Paso, Texas

PhD. In Data Science²
University of Texas at El Paso
El Paso, Texas

School of Theoretical and Applied Science³
Ramapo College of New Jersey
Mahwah, New Jersey

Transforming Time Series into Texture Images: A Fusion of Recurrence Plots and Gramian Angular Fields for Convolutional Neural Network-Based Classification

This paper presents a novel approach for converting time series data into texture images by combining recurrence plots and Gramian angular fields, termed as Recurrence-Gramian Convolutional Neural Network (RGCNN). Recurrence plots capture the temporal dependencies and recurrence patterns inherent in time series data, while Gramian angular fields encode the temporal dynamics into spatial structures resembling texture images. By combining these two techniques, we aim to create rich representations of time series data that can be effectively utilized by CNNs for classification tasks.

View Paper