Text Detection and Recognition Using Stroke Width Transform (SWT)

摘要

A technique is proposed to detect and recognize the texts. Localizing and reading text in uncontrolled environments remain extremely challenging, due to various interference factors. An integrated framework for text detection and recognition of varying natural images such as in vision applications, for instance, image understanding, image indexing, video search, geolocation and automatic navigation is proposed. The technique involves: Text detection and recognition using texts of varying orientation with dictionary search. Thus in this proposed system, a novel image database with text of different scales s colors, orientations is designed which can be detected and recognized synchronously. Low quality noisy images can also be included in the database.