Plataforma para desarrollo y evaluación de modelos de reconocimiento de gestos de la mano : modelo de reconocimiento de 11 gestos de la mano usando deep learning y celdas de memoria desarrollado en Python.

Abstract

Nowadays, technological advancements and their implementation in areas such as automation, virtual assistance, and human-machine interaction are constantly evolving. However, as artificial intelligence (AI) continues to expand its horizons, significant challenges related to model replicability in real-world settings also arise. The present research component focuses on the development of a Deep Learning model based on Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks for real-time classification and recognition of 11 hand gestures. The model utilizes electromyographic (EMG) and quaternion (IMU) signals as input. In this document, we delve into the process of creating a model using the CRISP-ML(Q) methodology for machine learning. Following each stage of this methodology, including signal acquisition, pre-processing, feature extraction, classification, and post-processing. Using this methodology, results were achieved, including a classification accuracy of 60.61% for the model and a recognition accuracy of 21.18%, for validation examples, using an application for evaluating hand gesture recognition models.