Naslov
Workflow for image categorization using cognitive computing services

Autori
Šeškar, Andrej ; Milašinović, Boris ; Fertalj, Krešimir

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Mipro 2018 proceedings / Skala, Karolj - Rijeka : Institute of Electrical and Electronics Engineers (IEEE), 2018, 1797-1802

ISBN
978-953-233-096-0

Skup
41st International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO 2018)

Mjesto i datum
Opatija, Hrvatska, 21.05.2018. - 25.05.2018

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
workflow ; image categorization ; Azure Logic Apps ; cognitive computing

Sažetak
The paper describes the structure of an Azure Logic App and gives a brief overview of the most relevant Computer Vision API services. Several solutions had been tested on a sample set consisting of vehicles photos in various street environments to show which of information can be extracted, e.g. licence plates, class of car, etc. and to estimate feasibility of different development scenarios. As an example of integration of various cloud services, a workflow for image categorization using Microsoft Computer Vision API and Azure Logic Apps had been developed as described in this paper. The Logic Apps workflow is triggered by taking a photo on a mobile phone and uploading it to a cloud storage for which a Logic App connector exists. The paper describes the developed workflow and discusses possibilities of replacing any component with another one, i.e. replacing a workflow trigger or the cognitive computing service. An approximate maintenance price calculation is done by estimating number of executions for each Logic Apps function based on a sample set of photos. The paper concludes with solution drawbacks and future work possibilities.

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo

POVEZANOST RADA