Comparing the Performance of Emotion-Recognition Implementations in OpenCV, Cognitive Services, and Google Vision APIs

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F70883521%3A28140%2F17%3A63517247" target="_blank" >RIV/70883521:28140/17:63517247 - isvavai.cz</a>

Výsledek na webu

<a href="http://www.wseas.org/multimedia/journals/information/2017/a405909-078.pdf" target="_blank" >http://www.wseas.org/multimedia/journals/information/2017/a405909-078.pdf</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Comparing the Performance of Emotion-Recognition Implementations in OpenCV, Cognitive Services, and Google Vision APIs

Popis výsledku v původním jazyce

Emotions represent feelings about people in several situations. Various machine learning algorithms have been developed for emotion detection in a multimedia element, such as an image or a video. These techniques can be measured by comparing their accuracy with a given dataset in order to determine which algorithm can be selected among others. This paper deals with the comparison of three implementations of emotion recognition in faces, each implemented with specific technology. OpenCV is an open-source library of functions and packages mostly used for computer-vision analysis and applications. Cognitive services, as well as Google Cloud AI, are sets of APIs which provide machine learning and artificial intelligence algorithms to develop smart applications capable of integrate computer-vision, speech, knowledge, and language processing features. Three Android mobile applications were developed in order to test the performance between an OpenCV algorithm for emotion recognition, an implementation of Emotion cognitive service, and a Google Cloud Vision deployment for emotion-detection in faces. For this research, one thousand tests were carried out per experiment. Our findings show that the OpenCV implementation got the best performance, which can be improved by increasing the sample size per emotion during the training step.

Název v anglickém jazyce

Comparing the Performance of Emotion-Recognition Implementations in OpenCV, Cognitive Services, and Google Vision APIs

Popis výsledku anglicky

Emotions represent feelings about people in several situations. Various machine learning algorithms have been developed for emotion detection in a multimedia element, such as an image or a video. These techniques can be measured by comparing their accuracy with a given dataset in order to determine which algorithm can be selected among others. This paper deals with the comparison of three implementations of emotion recognition in faces, each implemented with specific technology. OpenCV is an open-source library of functions and packages mostly used for computer-vision analysis and applications. Cognitive services, as well as Google Cloud AI, are sets of APIs which provide machine learning and artificial intelligence algorithms to develop smart applications capable of integrate computer-vision, speech, knowledge, and language processing features. Three Android mobile applications were developed in order to test the performance between an OpenCV algorithm for emotion recognition, an implementation of Emotion cognitive service, and a Google Cloud Vision deployment for emotion-detection in faces. For this research, one thousand tests were carried out per experiment. Our findings show that the OpenCV implementation got the best performance, which can be improved by increasing the sample size per emotion during the training step.

Druh

J_ost - Ostatní články v recenzovaných periodicích

CEP obor

—

OECD FORD obor

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

Projekt

Výsledek vznikl pri realizaci vícero projektů. Více informací v záložce Projekty.

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2017

Kód důvěrnosti údajů

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

Název periodika

WSEAS Transactions on Information Science and Applications

ISSN

1790-0832

e-ISSN

—

Svazek periodika

2017

Číslo periodika v rámci svazku

14

Stát vydavatele periodika

GR - Řecká republika

Počet stran výsledku

7

Strana od-do

184-190

Kód UT WoS článku

—

EID výsledku v databázi Scopus

—