Cursive Caption Text Detection in Videos
This work addresses the need for accurate text detection in videos for applications like semantic retrieval and opinion mining, though it is incremental as it adapts existing object detectors to a specific script.
The paper tackles the problem of detecting cursive caption text in video frames, specifically focusing on Urdu script, and achieves an F-measure of 0.91 on a dataset of around 11,000 frames.
Textual content appearing in videos represents an interesting index for semantic retrieval of videos (from archives), generation of alerts (live streams) as well as high level applications like opinion mining and content summarization. One of the key components of such systems is the detection of textual content in video frames and the same makes the subject of our present study. This paper presents a robust technique for detection of textual content appearing in video frames. More specifically we target text in cursive script taking Urdu text as a case study. Detection of textual regions in video frames is carried out by fine-tuning object detectors based on deep convolutional neural networks for the specific case of text detection. Since it is common to have videos with caption text in multiple-scripts, cursive text is distinguished from Latin text using a script-identification module. Finally, detection and script identification are combined in a single end-to-end trainable system. Experiments on a comprehensive dataset of around 11,000 video frames report an F-measure of 0.91.