Novel Method of Semantic Segmentation Applicable to Augmented Reality

• Published in: Sensors (Basel, Switzerland) Vol. 20; no. 6; p. 1737
• Main Authors: Ko, Tae-Young, Lee, Seung-Ho
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 20.03.2020; MDPI
• Subjects: Accuracy; Algorithms; atrous pyramid pooling module; Augmented reality; Back propagation; backpropagation; Classification; Convolution; convolutional neural network; Deep learning; Feature maps; Frames per second; fully convolutional network; Methods; modified dilated residual network; Neural networks; R&D; Research & development; Robots; Semantic segmentation; Semantics; Spatial data; modified dilated residual network; convolutional neural network; backpropagation; augmented reality; fully convolutional network; semantic segmentation; atrous pyramid pooling module

This paper proposes a novel method of semantic segmentation, consisting of modified dilated residual network, atrous pyramid pooling module, and backpropagation, that is applicable to augmented reality (AR). In the proposed method, the modified dilated residual network extracts a feature map from th...