Only a few works has been done for printed devanagari text in the area of optical character recognition. In this paper there is describing about a simple and fast algorithm for detection of italic and bold character in Devanagari script, without recognition of actual character. Here present an automatic information which tells us about the font type phase in the way of weight and slope. The pro...

This project aims to use machine learning as a way to make the optical character recognition of handwritten modern musical notation more accurate.

This paper demonstrates the use of segmentation, tracking, and number recognition of sport players for the provision of augmented information in moving sequences. Using prior knowledge of colour distribution the players and associated numbers are extracted for characterisation. The optical character recognition algorithms exploit multiple instances of the same character applying temporal filter...

When a document is fed to a scanner either mechanically or by a human operator for digitization, it suffers from some degrees of skew or tilt. Skew angle detection is an important component of any Optical Character Recognition (OCR) and Document Analysis System. In this paper we present a novel skew estimation technique for document images. The method uses Boundary Growing (BG) technique, Thinn...

In this paper , an improved skew angle estimation technique is presented. Skew angle detection is an important part of any Optical Character Recognition (OCR) and document analysis system. The method considers the lowermost pixels of some selected characters of the text, which maybe subject to Hough transform for skew angle detection. A fast approach is also proposed which works almost as accur...

This paper presents an automated system for authoring of hypermedia lecture recordings. We revised our previous work on lectures indexation. Our original method of lectures indexation needed a human interaction and was not feasible for large amount of lectures. We developed a system which utilizes Optical Character Recognition (OCR) to automatically extract the keywords from the lecture slides ...

In this paper we propose an algorithm which improves the quality of the digital version of antique books. In particular, we suggest a technique that enhances the virtual quality of the paper, eliminates some defects and improves the performance of Optical Character Recognition (OCR) operators.

Abstract: We demonstrate that a generative model for object shapes can achieve state of the art results on challenging scene text recognition tasks, and with orders of magnitude fewer training images than required for competing discriminative methods. In addition to transcribing text from challenging images, our method performs fine-grained instance segmentation of characters. We show that our ...

In this paper, we propose a method for correcting image distortions occurring when digitizing thick documents. The method corrects geometrical distortion using the document shape acquired by stereo measurement. Moreover, it obtains a fine restoration result by combining the better detailed portions in a stereo image pair. In an experiment, we show the validity of the method by investigating cha...

This paper gives an overview on the development process and ongoing research of the optical character recognition (OCR) systems for Telugu Text. The aim of this paper is to provide a starting point for the researchers entering into this field. In this paper, we present the introduction, characteristics of the Telugu language, developmental process of the OCR systems of Telugu language, research...