Improving web user interface element detection using Faster R-CNN

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F49777513%3A23520%2F21%3A43962817" target="_blank" >RIV/49777513:23520/21:43962817 - isvavai.cz</a>

Výsledek na webu

<a href="http://ceur-ws.org/Vol-2936/paper-117.pdf" target="_blank" >http://ceur-ws.org/Vol-2936/paper-117.pdf</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Improving web user interface element detection using Faster R-CNN

Popis výsledku v původním jazyce

Several challenges may arise when designing new user interfaces (UIs), e.g., because of communication between designers and developers, to which the detection of UI elements can help. The ImageCLEF DrawnUI 2021 challenge builds on the detection of such elements in two contest tasks: a Screenshot task that contains the website screenshot images with lots of noisy data, and a Wireframe task for detecting UI elements from hand-drawn proposals. This paper describes a simple algorithm based on the edge detection to filter noisy data from the website screenshots, and machine learning method which scored the first place in both tasks while having 0.628 and 0.900 mAP at 0.5 IoU in the Screenshot and Wireframe tasks. This method is based on the Faster R-CNN with a Feature Pyramid Network (FPN) that uses selected aspect ratios of anchor boxes according to the occurrences from the datasets. The code is available at https://github.com/vyskocj/ImageCLEFdrawnUI2021

Název v anglickém jazyce

Improving web user interface element detection using Faster R-CNN

Popis výsledku anglicky

Several challenges may arise when designing new user interfaces (UIs), e.g., because of communication between designers and developers, to which the detection of UI elements can help. The ImageCLEF DrawnUI 2021 challenge builds on the detection of such elements in two contest tasks: a Screenshot task that contains the website screenshot images with lots of noisy data, and a Wireframe task for detecting UI elements from hand-drawn proposals. This paper describes a simple algorithm based on the edge detection to filter noisy data from the website screenshots, and machine learning method which scored the first place in both tasks while having 0.628 and 0.900 mAP at 0.5 IoU in the Screenshot and Wireframe tasks. This method is based on the Faster R-CNN with a Feature Pyramid Network (FPN) that uses selected aspect ratios of anchor boxes according to the occurrences from the datasets. The code is available at https://github.com/vyskocj/ImageCLEFdrawnUI2021

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

20205 - Automation and control systems

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2021

Kód důvěrnosti údajů

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

Název statě ve sborníku

Proceedings of the Working Notes of CLEF 2021 - Conference and Labs of the Evaluation Forum

ISBN

—

ISSN

1613-0073

e-ISSN

—

Počet stran výsledku

12

Strana od-do

1375-1386

Název nakladatele

CEUR-WS

Místo vydání

Aachen

Místo konání akce

Bucharest, Romania(virtual)

Datum konání akce

21. 9. 2021

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—