Real-Time Turkish Video Text Detection and Recognition

Cet article n'a pas de fiche

Statut : exclu au tri — filtre mots-clés : aucun terme du groupe « origine ». Seuls les articles retenus au tri et dont le PDF est accessible sont lus en entier.

Résumé des auteurs

Detection and recognition of text in videos present significant challenges due to the wide range of font styles, varying text sizes, and diverse lighting conditions that can affect readability. The ability to accurately and efficiently detect text in such dynamic environments is essential for extracting meaningful information and enabling further processing. In this paper, real-time text detection and recognition for Turkish language has been analysed using different approaches. To identify the most suitable approach, multiple models, including EasyOCR, Tesseract, You Only Look Once, and Discrete Cosine Transform in conjunction with support vector machines were evaluated under different conditions. Since Turkish alphabet has similar letters, comparisons aim to improve the accuracy and speed of text extraction from video content, choosing a practical solution for real-time applications where precise text recognition is crucial.