Detection of Clickbait Thumbnails on YouTube Using Tesseract-OCR, Face Recognition, and Text Alteration

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Citation (Scopus)

Abstract

The combination of thumbnails and titles on the video-sharing website is effective for providing an overview of the content of the videos. But there are also clickbait thumbnails purposely designed to lure people into watching the videos, sometimes even misleading with the agenda of building a particular opinion. Clickbait thumbnails usually consist of a composition of photos of people and narrative text. OCR is generally used to detect text in images of similar objects, while face recognition is applied to identify people in photos. In this research, we implement OCR and face recognition to detect clickbait thumbnails. We use SVM Model to process the implementation results. Using dataset consist of 250 thumbnails, resulting in an accuracy value of 0.968, a sensitivity value of 0.968, a precision value of 0.9698, and an F1-Score of 0.9678.

Original languageEnglish
Title of host publicationICAICST 2021 - 2021 International Conference on Artificial Intelligence and Computer Science Technology
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages56-61
Number of pages6
ISBN (Electronic)9781665424042
DOIs
Publication statusPublished - 29 Jun 2021
Event2021 International Conference on Artificial Intelligence and Computer Science Technology, ICAICST 2021 - Virtual, Online
Duration: 29 Jun 2021 → …

Publication series

NameICAICST 2021 - 2021 International Conference on Artificial Intelligence and Computer Science Technology

Conference

Conference2021 International Conference on Artificial Intelligence and Computer Science Technology, ICAICST 2021
CityVirtual, Online
Period29/06/21 → …

Keywords

  • Face Recognition
  • OCR
  • Tesseract
  • Text Alteration
  • YouTube

Fingerprint

Dive into the research topics of 'Detection of Clickbait Thumbnails on YouTube Using Tesseract-OCR, Face Recognition, and Text Alteration'. Together they form a unique fingerprint.

Cite this