International Conference on Computer and Knowledge Engineering

Home / 12th International Conference on Computer and Knowledge Engineering

A 2D-CNN Architecture for Improving the Classification Accuracy of an Electronic Nose with Different Sensor Positions

Authors :

Hannaneh Mahdavi¹ Reza Goldoust² Saeideh Rahbarpour³

1- Department of Electrical Engineering Shahed University Tehran, Iran 2- Department of Electrical Engineering Shahed University Tehran, Iran 3- Department of Electrical Engineering Shahed University Tehran, Iran

Keywords :

Convolutional Neural Network (CNN)،Electronic Nose (E-Nose)،Feature extraction،Metal oxide gas sensor،Spatial information

Abstract :

The responses of Metal oxide gas sensors (MOXs) are affected by various factors; one of them is their location. For achieving a good classification accuracy by an Electronic Nose (E-Nose), extracting informative features with consideration of the spatial information of signals is necessary. A popular E-Nose dataset consisting of the responses of 72 MOX sensors from eight types and in nine positions to 10 pollutants in 1165 experiments was used to investigate the importance of considering the location of sensors. A method is proposed based on a simple Two-dimensional Convolutional Neural Network (2D-CNN) and compared to a 1D-CNN with the same number of parameters. It is shown that the 2D-CNN scheme results in 97.8% detection accuracy, which is 7.5% upper than the accuracy value of 1D-CNN. It is concluded that by considering the spatial and temporal information of signals by 2D-CNN, better feature extraction and a more accurate classifier could be reached.