On the relevance of the metadata used in the semantic segmentation of indoor image spaces

https://doi.org/10.1016/j.eswa.2021.115486Get rights and content
Under a Creative Commons license
open access

Highlights

  • Quantitatively evaluate the usefulness of contextual information for a U-Net.

  • Prove that learning systems rely heavily on contextual info for identification tasks.

  • The Importance of Metadata Applied to Semantic Segmentation for Indoor Scenes.

  • Deploying an efficient and robust FCN for Semantic Segmentation of Indoor Imagery.

Abstract

The study of artificial learning processes in the area of computer vision context has mainly focused on achieving a fixed output target rather than on identifying the underlying processes as a means to develop solutions capable of performing as good as or better than the human brain. This work reviews the well-known segmentation efforts in computer vision. However, our primary focus is on the quantitative evaluation of the amount of contextual information provided to the neural network. In particular, the information used to mimic the tacit information that a human is capable of using, like a sense of unambiguous order and the capability of improving its estimation by complementing already learned information. Our results show that, after a set of pre and post-processing methods applied to both the training data and the neural network architecture, the predictions made were drastically closer to the expected output in comparison to the cases where no contextual additions were provided. Our results provide evidence that learning systems strongly rely on contextual information for the identification task process.

Keywords

Deep learning
U-net
Semantic segmentation
Metadata preprocessing
Fully convolutional network
Indoor scenes

Cited by (0)

Luis Vasquez-Espinoza, graduated of Computer Science program at Universidad Nacional de Ingenieria (Peru), has been working as a senior investigator at the computing specialized scientific association (ACECOM-UNI) and at Intelligent Ubiquitous Technologies – Smart City (IUT-SCi) Lab. His current research interests include optimization of deep learning architectures and evaluating the similarities with the behavior of the human brain.

Manuel Castillo-Cara received the PhD degree from the University of Castilla-La Mancha (Spain) in July 2018. He has been working on university educational issues at the Computer Science as an Associate Professor and head of Intelligent Ubiquitous Technologies – Smart City (IUT-SCi) Lab at Universidad de Lima (Peru). His current research is focused on Intelligent Ubiquitous Technologies, especially on in Wireless Sensor Networks, Distributed Computing, Pattern Recognition and Artificial Intelligence.

Luis Orozco-Barbosa received the Doctorat de l’Université from the Université e Pierre et Marie Curie, France, in 1987. From 1987 to 2001, he was a faculty member at the Electrical and Computer Engineering Department of the University of Ottawa, Canada. In 2002, he joined the Department of Computer Engineering at the Universidad de Castilla-La Mancha (Spain). His current research interests include Internet protocols, wireless sensor communications, and IoT technologies. He is a member of the IEEE.