Paper Title
Scene Text Recognition Via Part Based Tree Structured Model

Abstract- Scene text recognition has encouraged great interests from the computer vision area in recent years. A novel scene text recognition method which converts the word into speech is proposed. A part based tree structure is used to model characters so as to detect and recognize characters. The character models make use of both the local appearance and global structure information. And thus the detection results are more reliable. And these detected words are converted into speech using Microsoft SAPI. Keyword- Ocr, Speech Synthesizer, Sapi, Text Recognition.