Analysis and Modeling of Alcohol Intoxication from IR Images based on Multiregional Image Segmentation and Correlation with Breath Analysis

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F61989100%3A27240%2F17%3A10238646" target="_blank" >RIV/61989100:27240/17:10238646 - isvavai.cz</a>

Výsledek na webu

<a href="http://ieeexplore.ieee.org/document/8284106/" target="_blank" >http://ieeexplore.ieee.org/document/8284106/</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/ICBDAA.2017.8284106" target="_blank" >10.1109/ICBDAA.2017.8284106</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Analysis and Modeling of Alcohol Intoxication from IR Images based on Multiregional Image Segmentation and Correlation with Breath Analysis

Popis výsledku v původním jazyce

Alcohol intoxication is an important procedure which is related to all social areas. There are commonly used standards like are the blood analysis or breath analysis. Although such methods are commonly used and give satisfactory results, there are also certain drawbacks. For instance it is direct contact and awareness of the tested person. One challenging direction of the alcohol assessment is the temperature effect whilst drinking, thus temperature variations may be reliable indicators of the current alcohol state. The paper deals with a comparative analysis of three multiregional segmentation methods with target of building of a mathematical model reflecting the facial areas well reflecting the dynamical process of the alcohol drinking. By such modelling we can make a predictor of the alcohol state based on the facial temperature effect. We have specified two significant features: nose and forehead areas where the temperature variations are well observable. Eventually, we have done a verification analysis between individual dynamical models and breath analysis on the base of the Pearson correlation coefficient. Correlation gives relatively strong dependence, when we consider a fact that some persons have stronger inclination to the alcohol which may negatively influence the IR records and the segmentation results as well.

Název v anglickém jazyce

Analysis and Modeling of Alcohol Intoxication from IR Images based on Multiregional Image Segmentation and Correlation with Breath Analysis

Popis výsledku anglicky

Alcohol intoxication is an important procedure which is related to all social areas. There are commonly used standards like are the blood analysis or breath analysis. Although such methods are commonly used and give satisfactory results, there are also certain drawbacks. For instance it is direct contact and awareness of the tested person. One challenging direction of the alcohol assessment is the temperature effect whilst drinking, thus temperature variations may be reliable indicators of the current alcohol state. The paper deals with a comparative analysis of three multiregional segmentation methods with target of building of a mathematical model reflecting the facial areas well reflecting the dynamical process of the alcohol drinking. By such modelling we can make a predictor of the alcohol state based on the facial temperature effect. We have specified two significant features: nose and forehead areas where the temperature variations are well observable. Eventually, we have done a verification analysis between individual dynamical models and breath analysis on the base of the Pearson correlation coefficient. Correlation gives relatively strong dependence, when we consider a fact that some persons have stronger inclination to the alcohol which may negatively influence the IR records and the segmentation results as well.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2017

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Big Data and Analytics (ICBDA) : conference proceedings : November 16-17, 2017, Kuching, Malaysia

ISBN

978-1-5386-0790-9

ISSN

—

e-ISSN

neuvedeno

Počet stran výsledku

5

Strana od-do

49-54

Název nakladatele

IEEE

Místo vydání

Piscataway

Místo konání akce

Kuching

Datum konání akce

16. 11. 2017

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

000426452100009