Analysis of pattern recognition and dimensionality reduction techniques for odor biometrics

Abstract

In this paper, we analyze the performance of several well-known pattern recognition and dimensionality reduction techniques when applied to mass-spectrometry data for odor biometric identification. Moti-vated by the successful results of previous works capturing the odor from other parts of the body, this work attempts to evaluate the feasibility of identifying people by the odor emanated from the hands. By formulating this task according to a machine learning scheme, the problem is identified with a small-sample-size supervised classification problem in which the input data is formed by mass spectro-grams from the hand odor of 13 subjects captured in different sessions. The high dimensionality of the data makes it necessary to apply feature selection and extraction techniques together with a simple clas-sifier in order to improve the generalization capabilities of the model. Our experimental results achieve recognition rates over 85% which reveals that there exists discriminatory information in the hand odor and points at body odor as a promising biometric identifier.

State of the art and perspectives on traditional and emerging biometrics: A survey

“…Rodriguez‐Lujan et al investigated the possibility of using the body odor for biometric identification. They conducted an experiment in which they captured 728 samples from 13 subjects over 28 sessions.…”

Odorant recognition using biological responses recorded in olfactory bulb of rats

“…It is interesting to point out that LDA can be used either as a dimensionality reduction technique or as a classifier. In our case we will use it as a classifier as it is used in [53]. These classifiers are briefly described in what follows.…”

Recent Developments in Human Odor Detection Technologies

“…Detection of human odor has become a subject of significant research interest because of the rich information on human body chemistry that can be extracted from human odor and the wide variety of potential applications this opens up. Since each person's body produces unique odors regardless of diet, human odor can be used as a biometric for non-invasive and non-collaborative subject identification [1][2][3]. The tracking of subject location as well as the path or objects in which the subject comes into contact via human odor detection is also a potential application for criminal investigation and forensics [4,5].…”