Visual recognition incorporating features of self-supervised models for the use of unlabelled data

Resultado de la investigación: Contribución a los tipos de informe/libroContribución a la conferenciarevisión exhaustiva

Resumen

Automatic visual object recognition has gained great popularity in the world and is successfully applied to various areas such as robotics, security or commerce using deep learning techniques. Training in machine learning models based on deep learning requires an enormous amount of supervised data, which is expensive to obtain. An alternative is to use semi-supervised models as co-training where the views given by deep networks are differentiated using models that incorporate lateral information from each training object. In this document, we describe and test a co-training model for deep networks, adding as auxiliary inputs to self-supervised network features. The results show that the proposed model managed to converge using a few dozen iterations, exceeding 2 % in precision compared to recent models. This model, despite its simplicity, manages to be competitive with more complex recent works. As future work, we plan to modify deep self-supervised networks to increase diversity in co-training learning.

Idioma originalInglés
Título de la publicación alojada2021 IEEE International Conference on Automation/24th Congress of the Chilean Association of Automatic Control, ICA-ACCA 2021
EditorialInstitute of Electrical and Electronics Engineers Inc.
ISBN (versión digital)9781665401272
DOI
EstadoPublicada - 22 mar 2021
Evento2021 IEEE International Conference on Automation/24th Congress of the Chilean Association of Automatic Control, ICA-ACCA 2021 - Valparaiso, Chile
Duración: 22 mar 202126 mar 2021

Serie de la publicación

Nombre2021 IEEE International Conference on Automation/24th Congress of the Chilean Association of Automatic Control, ICA-ACCA 2021

Conferencia

Conferencia2021 IEEE International Conference on Automation/24th Congress of the Chilean Association of Automatic Control, ICA-ACCA 2021
País/TerritorioChile
CiudadValparaiso
Período22/03/2126/03/21

Áreas temáticas de ASJC Scopus

  • Informática aplicada
  • Seguridad, riesgos, fiabilidad y calidad
  • Control y optimización
  • Informática aplicada a la salud
  • Redes de ordenadores y comunicaciones

Huella

Profundice en los temas de investigación de 'Visual recognition incorporating features of self-supervised models for the use of unlabelled data'. En conjunto forman una huella única.

Citar esto