Text extraction from natural scene image: A survey

In recent years, with the widespread of Internet and digitized processing of multi-script documents worldwide, script identification techniques have become more important in the pattern recognition field. Script identification concerns methods for identifying different scripts in multi-lingual, multi-script documents. This paper presents a comprehensive overview on research activities in the field and focuses on the most valuable results obtained so far. The most vital processes in script identification are addressed in detail: identification and discriminating methods, features extraction (local and global), and classification. Different kinds of approaches have been developed and promising results have been achieved. This paper reports SoA performance results. This paper reports methods concerning handwritten, printed, and hybrid document processing. More research is necessary to meet the performance levels essential for everyday applications.

Script Identification of Multi-Script Documents: A Survey

Wireless electric appliance control for smart buildings using indoor location tracking and BIM-based virtual environments

Offline Script Identification from multilingual Indic-script documents: A state-of-the-art

Natural scenes of blurred text images are challenges to current text recognition field. In our paper, a novel method for text detection in natural scene image is suggested using edge detection, maximally stable extremal region (MSER) and tensor voting. Edge detection and MSER methods are combined to find the greatest character candidates from stable areas which are extracted from an input image. These text candidates are used to extract the text line information using tensor voting that creates normal vectors and curve saliency values in characters along the text lines. Therefore, the text line information is used to eliminate non-text areas. Our method is evaluated on the ICDAR2013 datasets and experiment results show that the proposed result is compared to the previous methods.

Robust Text Detection in Natural Scene Images

Automated systems for understanding display boards are finding many applications useful in guiding tourists, assisting visually challenged and also in providing location aware information. Such systems require an automated method to detect and extract text prior to further image analysis. In this paper, a methodology to detect and extract text regions from low resolution natural scene images is presented. The proposed work is texture based and uses DCT based high pass filter to remove constant background. The texture features are then obtained on every 50×50 block of the processed image and potential text blocks are identified using newly defined discriminant functions. Further, the detected text blocks are merged and refined to extract text regions. The proposed method is robust and achieves a detection rate of 96.6% on a variety of 100 low resolution natural scene images each of size 240×320.

Text region extraction from low resolution natural scene images using texture features

License Plate Recognition (LPR) has been intensively studied in many countries. Due to the different types of number plates being used, the requirements of an automatic number plate recognition system are different for each country. In general, objective of any LPR system is to localize and recognize potential license plate region(s) of the vehicle images captured from handheld device/digital camera/mobile phone camera. The proposed method for License Plate Recognition (LPR) system works in three modules: localization of license plate, segmentation of the characters and recognition of the characters from the license plate. Localization of the license plate is done using morphological operations, horizontal & vertical edge processing. Character segmentation is carried out using connected component labeling. Character recognition is done by using Neural Network classifier. The method is tested on 100 samples of Indian vehicle images. The system achieves 86% accuracy in localization, 81% accuracy in segmentation and 80% accuracy in character recognition.

License plate recognition system for indian vehicles

Automated systems for understanding low resolution images of display boards are facilitating several new applications such as blind assistants, tour guide systems, location aware systems and many more. Script identification at character/word level is one of the very important pre-processing steps for development of such systems prior to further image analysis. In this paper, a new fuzzy based approach for word level script identification of text in low resolution images of display boards is presented. The proposed methodology uses horizontal run statistics and wavelet features for distinguishing 5 Indian scripts namely; Hindi, Kannada, English, Malyalam and Tamil. The method works in two phases; In the first phase, the wavelet transform based texture features such as zone wise wavelet energy features, vertical run statistical features of wavelet coefficients and wavelet log mean deviation features of decomposed energy bands at 2 levels are obtained from training word images and crisp sets are constructed, one for each script/language under study. The second phase is testing, in which test word image is processed to obtain horizontal run statistics to determine whether it belongs to Hindi script. Otherwise, the word image is processed to obtain a crisp vector. The degree of belongingness of crisp vector with each candidate object in the crisp sets is determined using newly devised fuzzy membership function. Further, fuzzy inference scheme is used to identify the script of the test word image. The proposed method is robust and insensitive to the variations in size and style of font, number of characters, thickness and spacing between characters, noise, and other degradations. The proposed method achieves an overall script identification accuracy of 94.33% and individual identification accuracy of 100% for Hindi Script, 98.67% for Kannada Script, 100% for English, 89% for Malyalam and 84% for Tamil Script.

A fuzzy approach for word level script identification of text in low resolution display board images using wavelet features

Internet of Things (IoT) is a technology which enables computing devices, physical and virtual objects/devices to be connected to the internet so that users can control and monitor devices. The IoT offers huge potential for development of various applications namely: e-governance, environmental monitoring, military applications, infrastructure management, industrial applications, energy management, healthcare monitoring, home automation and transport systems. In this paper, brief overview of existing frameworks for development of IoT applications, techniques to develop smart home applications using existing IoT frameworks and generic framework for the development of IoT based smart home system is presented. The proposed generic framework comprises various modules such as Auto-Configuration and Management, Communication Protocol, Auto-Monitoring and Control, and Objects Access Control. The architecture of the new generic framework and the functionality of various modules in the framework are also presented. The proposed generic framework is helpful for making every house as smart house to increase the comfort of inhabitants. Each of the components of generic framework is robust in nature in providing services at any time. The components of smart home system are designed to take care of various issues such as scalability, interoperability, device adaptability, security and privacy. The proposed generic framework is designed to work on all vendor boards and variants of Linux and Windows operating system.

A review on home automation system

Reliable extraction/segmentation of text lines, words and characters is one of the very important steps for development of automated systems for understanding the text in low resolution display board images. In this paper, a new approach for segmentation of text lines, words and characters from Kannada text in low resolution display board images is presented. The proposed method uses projection profile features and on pixel distribution statistics for segmentation of text lines. The method also detects text lines containing consonant modifiers and merges them with corresponding text lines, and efficiently separates overlapped text lines as well. The character extraction process computes character boundaries using vertical profile features for extracting character images from every text line. Further, the word segmentation process uses k-means clustering to group inter character gaps into character and word cluster spaces, which are used to compute thresholds for extracting words. The method also takes care of variations in character and word gaps. The proposed methodology is evaluated on a data set of 1008 low resolution images of display boards containing Kannada text captured from 2 mega pixel cameras on mobile phones at various sizes 240 × 320, 480 × 640 and 960 × 1280. The method achieves text line segmentation accuracy of 97.17%, word segmentation accuracy of 97.54% and character extraction accuracy of 99.09%. The proposed method is tolerant to font variability, spacing variations between characters and words, absence of free segmentation path due to consonant and vowel modifiers, noise and other degradations. The experimentation with images containing overlapped text lines has given promising results.

M. M. Kodabagi

Papers

Text region extraction from low resolution natural scene images using texture features

License plate recognition system for indian vehicles

A fuzzy approach for word level script identification of text in low resolution display board images using wavelet features

A review on home automation system

A Robust Segmentation Technique for Line, Word and Character Extraction from Kannada Text in Low Resolution Display Board Images