Дослідження та розробка системи автоматичного розпізнавання мовлення та генерації субтитрів для веб-платформ

Abstract

The paper presents the results of a study on the automation of subtitle generation for short-form video content using automatic speech recognition (ASR). The relevance of the problem is substantiated by the growing role of subtitled video in social media marketing and the limitations of existing manual approaches. An asynchronous microservice architecture for a web-based subtitle generation platform is proposed, enabling a complete processing pipeline from video upload to the production of a ready-to-publish file with embedded subtitles. The feasibility and efficiency of the proposed solution are confirmed through the development and experimental evaluation of a functional prototype.