My name is sanjay Mali resident at indore mp what is my id
Posts: 14,118
Threads: 61
Joined: Oct 2014
The use of OCR would be the perfect solution in terms of results and programming. However, you can try to develop your own algorithm using object detection and recognition techniques. Use the Bounding Box statement in MATLAB to detect all objects and compare objects with existing images of all alphabets. To compare, you can use co-efficient or mutual co-reporting. The former is computationally expensive and would require programming skills.
The text data present in images and video contain useful information for automatic annotation, indexing and image structuring. The extraction of this information involves the detection, localization, tracking, extraction, improvement and recognition of the text of a given image. However, text variations due to differences in size, style, orientation and alignment, as well as under complex image and background contrast make the problem of automatic text extraction extremely difficult. While comprehensive surveys of related issues such as face detection, document analysis, and indexing of images and videos can be found, the problem of extracting text information is not well studied. A number of techniques have been proposed to address this problem and the purpose of this paper is to classify and revise these algorithms, discuss baseline data and performance assessment, and point out promising directions for future research.