The traditional test extraction suffers from the drawback of size style and rotation of text arose on the images. Thus the scanning device needs to focus on the textual region of the images. Which is going to involve the person who is using the application software? This can be automated using algorithm which is written and designed in such a way so that the text area on the image will be easily identified either at some orientation or at variable sizes. In the presented work, the characters are segmented by using the pixel neighborhood technique and resized to a 32x32 block.
The centre of gravity of the different characters is computed by using the first order moments. The contour of the pixel is extracted by means of Robert’s operator. The radii from centre of gravity to contour pixel and are arranged in descending order. If the same character is rotated about its centre of gravity by some angle, the same radii are extracted and are arranged in descending order.
It is observed that the first few radii are same for the same character if rotated at any angle. This gives the rotation invariant character recognition. Further, the characters are normalized with respect to size by dividing the radii by mean radius. The location invariance is obtained by use of centre of gravity. In the proposed algorithm, the different in-variances are considered into the features extraction process such that the normalization of characters is done in all respect. Once the features of different characters are set and are constant for the same object in either form, then that features can be used for character identification purposes.
Table of Contents
CHAPTER 1
INTRODUCTION
1.1 ACCURACY WITH OCR
1.2 BACK-PROPAGATION NEURAL NETWORK CLASSIFIER
1.2.1 First Phase: Propagation
1.2.2 Second Phase: Weight Update
1.3 ALGORITHM
CHAPTER 2
LITERATURE REVIEW
CHAPTER 3
PRESENT WORK
3.1 PROBLEM FORMULATION
3.2 OBJECTIVES
3.3 IMAGE PRE-PROCESSING
3.4 CHARACTER SEGMENTATION AND NORMALIZATION
3.5 CHARACTER IDENTIFICATION
3.6 APPLICATION CHARACTER RECOGNITION SYSTEM
3.6.1. Text/Characters Segmentation and Training:
3.7 HARDWARE AND SOFTWARE REQUIREMENTS
CHAPTER 4
RESULTS AND DISCUSSION
CHAPTER 5
CONCLUSION AND FUTURE SCOPE
5.1 CONCLUSION
5.2 FUTURE SCOPE
Objectives and Research Focus
This work aims to develop an efficient, automated system for text extraction and character recognition from documented images, addressing challenges related to variable image orientation, size, and location. By designing an algorithm that utilizes a neural network trained on statistical character features, the research seeks to minimize manual data entry and create a reliable workflow for converting digital document images into editable text files.
- Analysis of existing image segmentation and partitioning techniques.
- Development of a robust neural network-based algorithm for text extraction.
- Implementation of invariant normalization processes for character features.
- Performance evaluation based on training accuracy and epoch-based optimization.
Excerpt from the Book
INTRODUCTION
Extraction of text from documented images finds application in maximum entries which are document related in offices. The most of the popular applications which we find in public or college libraries where the entries of number of books are done by manually typing the title of book along with other credentials like name of the author and other attributes. The complete process can be made effortless with the application of a suitable algorithm or application software which can be extract the documented part from the cover of book and other parts of the book thereby reducing the manual job like typing of user. Which reduces the overall job to only arranging the book title etc.by formatting the material. [1]
The goal of document image analysis is to covert the documented information hidden on a digitized images into a well scripted symbolic representation. The main information carriers are the textual parts in most of the applications. Hence it is important to locate text blocks within the images, recognize the text and extract the hidden documents. The documents consists of texts, graphics and images which may overlap. Since the alignment of text may not be always horizontally aligned thus finding the documented images and segmenting words, characters and lines are not a trivial task. Due to the huge reduction in the memory space of the concluded results, it is fruitful to generate transmit and save the document in the reevaluated form. The region which is extracted can be processed by number of steps depending on their types e.g. OCR for documented images blocks and compression for graphical images together with halftone images. Up to now several strategies have been tried to solve the problem of segmentation. [2]
Summary of Chapters
CHAPTER 1: Provides an introduction to the need for automated text extraction from images and outlines the fundamental concepts of OCR and back-propagation neural networks.
CHAPTER 2: Reviews existing research and literature concerning image analysis, character recognition, and various neural network approaches to pattern matching.
CHAPTER 3: Details the proposed methodology, including image pre-processing, character segmentation, feature extraction, and the specific architecture of the neural network developed.
CHAPTER 4: Presents the experimental results, including training configurations, accuracy metrics, and visual analysis of the system's performance across various epochs.
CHAPTER 5: Concludes the study by summarizing the effectiveness of the algorithm and discusses potential future applications for the developed system.
Keywords
Optical Character Recognition, OCR, Back-propagation, Neural Network, Image Segmentation, Document Analysis, Feature Extraction, Pattern Recognition, Binarization, Normalization, Text Extraction, Algorithm, Matlab, Statistical Features, Digitization
Frequently Asked Questions
What is the core focus of this research?
The research focuses on the automated extraction and recognition of text from document images using neural networks, aiming to reduce manual labor in data entry tasks.
What are the primary thematic fields covered in this work?
The work covers image processing, document analysis, machine learning through back-propagation neural networks, and feature extraction techniques.
What is the primary goal of the study?
The primary goal is to design an algorithm that successfully recognizes characters from images regardless of their original orientation, size, or location.
Which scientific method is employed for the recognition process?
The study utilizes a back-propagation neural network classifier, which is trained on statistical feature sets extracted from character images.
What topics are discussed in the main part of the book?
The main part covers the problem formulation, image pre-processing (binarization), feature extraction methodologies, neural network training phases, and hardware/software implementation requirements.
Which keywords define this work?
Key terms include Optical Character Recognition (OCR), neural networks, image segmentation, pattern recognition, and feature normalization.
How does the algorithm handle different font sizes and orientations?
The algorithm utilizes a normalization process that ensures features are independent of size, location, and rotation, allowing the system to remain consistent across different character styles.
What is the significance of the "centre of gravity" (COG) in this research?
The COG is computed during the character segmentation phase as a reference point for statistical moments, helping to ensure the extracted features are invariant to the character's location in the image.
How is the accuracy of the neural network validated?
The accuracy is validated by training the network with 180 sets of characters and subsequently testing it with varied inputs, achieving an accuracy rate above 97%.
- Quote paper
- Pankaj Bhambri (Author), 2018, Inventing a Recognition System to Rotate, Scale and Translate Invariant Characters, Munich, GRIN Verlag, https://www.grin.com/document/427526
