Two-dimensional wavelet transform feature extraction for porous silicon chemical sensors

Area of Science:

• Chemical Sensing
• Materials Science
• Signal Processing

Background:

• Achieving reliable, sensitive, fast, and low-power chemo-sensory systems remains a significant challenge.
• Existing chemo-sensory detectors often fall short of ideal performance criteria.
• Porous silicon (pSi) based optical gas sensors offer potential but require advanced analysis techniques.

Purpose of the Study:

• To explore physicochemical interactions in chemically modified pSi film-based optical gas sensors.
• To introduce and evaluate the two-dimensional discrete wavelet transform (2D-DWT) as a feature extraction method.
• To improve the discrimination and quantification of chemical analytes using optical sensor data.

Main Methods:

• Implementation of the 2D-DWT to analyze the bi-dimensional response of pSi rugate sensors.
• Utilizing signal processing to capture non-stationary behavior in sensor responses.
• Applying the method to a six-dimensional chemical analyte discrimination and quantification task.

Main Results:

• The 2D-DWT effectively extracts features from the bi-dimensional optical sensor response.
• The proposed method demonstrates significance in a complex chemical analyte analysis.
• Evidence suggests the 2D-DWT enhances understanding of optically based chemical sensor performance.

Conclusions:

• The 2D-DWT is a valuable tool for analyzing pSi optical gas sensors.
• This feature extraction strategy deepens insights into sensor-analyte interactions.
• The approach represents a step towards more realistic chemo-sensing applications.