Text Extraction Method Research Articles

Automated Extraction of Large Scale Scanned Document Images using Google Vision OCR in Apache Hadoop Environment

This Digitalization of documents is now being done in all fields to reduce paper usage. The availability of modern technology in the form of scanners and cameras supports the growth of multimedia data, especially documents stored in the form of image files. Searching a particular text in a large-scale scanned document images is a difficult task if the document is in the form of images where the text has not been extracted. In this research, text extraction method of large-scale scanned document images using Google Vision OCR on the Hadoop architecture is proposed. The object of research is student thesis documents, which includes the cover page, the approval page, and abstract. All documents are stored in the university's digital library. Extraction process begins with preparing the input folder that contains image documents (in JPEG format) in HDFS Apache Hadoop and followed by reading the image document. The image document is then extracted using Google Vision OCR in order to obtain text document (in TXT format) and the result is saved to output folder in Hadoop Distributed File System (HDFS). The same process is repeated for the entire documents in the folder. Test results have shown that the proposed methods was able to extract all test documents successfully. The recognition process achieved 100% accuracy and the extraction time is twice as fast as manual extraction. Google Vision OCR also shows better extraction performance compared to other OCR tools. The proposed automated extraction systems can recognize text in a large-scale image document accurately and can be operated in a real-time environment.

A Novel M-ACA-Based Tumor Segmentation and DAPP Feature Extraction with PPCSO-PKC-Based MRI Classification

In medical image processing, segmentation and classification plays a significant part in prediction of affected region from given input image. The novel segmentation and classification model to segment tumor region and to identify the abnormality category from MR brain image is projected. In preprocessing, image filtering by distribution-based adaptive median filtering technique to prove smoothness to the image and to eliminate the noise component is provided. Further the skull region is removed by using adaptive threshold-based edge detection with canny method. In the segmentation, a novel multiangle cellular automata model to predict the region of interest, i.e., tumor spot is discussed. The classification performance is improved by novel texture extraction and optimal feature selection method named as dynamic angle projection pattern and priority particle cuckoo search optimization, respectively. These optimized features are given to support vector machine (SVM) and pointing kernel classifier (PKC) to classify the abnormality level of segmented image. This work can be compared with existing systems for the parameters of sensitivity, specificity, accuracy, FPR, TPR, and ROC. Our proposed classifier PKC’s performance is good when compared with other technique. The sensitivity, specificity, and accuracy of 85.9155, 94.3396, 89.5161, and 95.7746%, 100, and 97.5806% are obtained in SVM and PKC, respectively. SVMs are linear up to 0.8 for higher values of FPR and reaches the stable point prior to the maximum value. But, the stable operation is achieved in proposed PKC at 0.9. Result shows that the improvement in performance of proposed PKC.

Text extraction method for historical Tibetan document images based on block projections

Text extraction is an important initial step in digitizing the historical documents. In this paper, we present a text extraction method for historical Tibetan document images based on block projections. The task of text extraction is considered as text area detection and location problem. The images are divided equally into blocks and the blocks are filtered by the information of the categories of connected components and corner point density. By analyzing the filtered blocks’ projections, the approximate text areas can be located, and the text regions are extracted. Experiments on the dataset of historical Tibetan documents demonstrate the effectiveness of the proposed method.

Evaluation of effectiveness of three fuzzy systems and three texture extraction methods for building damage detection from post-event LiDAR data

ABSTRACTBuilding damage maps after disasters can help us to better manage the rescue operations. Researchers have used Light Detection and Ranging (LiDAR) data for extracting the building damage maps. For producing building damage maps from LiDAR data in a rapid manner, it is necessary to understand the effectiveness of features and classifiers. However, there is no comprehensive study on the performance of features and classifiers in identifying damaged areas. In this study, the effectiveness of three texture extraction methods and three fuzzy systems for producing the building damage maps was investigated. In the proposed method, at first, a pre-processing stage was utilized to apply essential processes on post-event LiDAR data. Second, textural features were extracted from the pre-processed LiDAR data. Third, fuzzy inference systems were generated to make a relation between the extracted textural features of buildings and their damage extents. The proposed method was tested across three areas over the 2010 Haiti earthquake. Three building damage maps with overall accuracies of 75.0%, 78.1% and 61.4% were achieved. Based on outcomes, the fuzzy inference systems were stronger than random forest, bagging, boosting and support vector machine classifiers for detecting damaged buildings.

Text Extraction and Recognition from the Normal Images using MSER Feature Extraction and Text Segmentation Methods

Image mining is concerned with the extraction of contained information, image information connection or other patterns not clearly stored in the images. Text in images is one of the dominant features and its extraction is a big task. If this type of text could be segmented, detected, extracted and recognized automatically, than it would be a precious source of high-level retrieval process. In the research work, text extraction and recognition from the normal images using MSER feature extraction and text segmentation methods has been developed to detect the text regions and the system is based on efficient optical character recognition process. Text extraction and recognition from the normal images is important for content based image analysis. This problem is challenging due to the complex background of images, reflection of light in images and shadow portion presented in images. The proposed technique in this work develops a well-organized text extraction and recognition methods that utilizes the concept of morphological operations using digital image processing. Existing text extraction method, namely, region based method produces enhanced results when applied on the normal images. The advantage of segmentation for the feature extraction of text region is proposed in the system.

A Fuzzy-GA Based Decision Making System for Detecting Damaged Buildings from High-Spatial Resolution Optical Images

In this research, a semi-automated building damage detection system is addressed under the umbrella of high-spatial resolution remotely sensed images. The aim of this study was to develop a semi-automated fuzzy decision making system using Genetic Algorithm (GA). Our proposed system contains four main stages. In the first stage, post-event optical images were pre-processed. In the second stage, textural features were extracted from the pre-processed post-event optical images using Haralick texture extraction method. Afterwards, in the third stage, a semi-automated Fuzzy-GA (Fuzzy Genetic Algorithm) decision making system was used to identify damaged buildings from the extracted texture features. In the fourth stage, a comprehensive sensitivity analysis was performed to achieve parameters of GA leading to more accurate results. Finally, the accuracy of results was assessed using check and test samples. The proposed system was tested over the 2010 Haiti earthquake (Area 1 and Area 2) and the 2003 Bam earthquake (Area 3). The proposed system resulted in overall accuracies of 76.88 ± 1.22%, 65.43 ± 0.29%, and 90.96 ± 0.15% over Area 1, Area 2, and Area 3, respectively. On the one hand, based on the concept of the proposed Fuzzy-GA decision making system, the automation level of this system is higher than other existing systems. On the other hand, based on the accuracy of our proposed system and four advanced machine learning techniques, i.e., bagging, boosting, random forests, and support vector machine, in the detection of damaged buildings, it seems that our proposed system is robust and efficient.

Flotation froth image texture extraction method based on deterministic tourist walks

In the flotation process, the flotation froth texture is an indicator of the flotation state. To recognize the flotation state based on texture features accurately and to provide guidance for production operations, this paper proposes a method for flotation froth image texture extraction based on the deterministic tourist walks algorithm. First, a weighted graph model of a froth image is built using deterministic tourist walks. Next, the degree distribution and the unit intensity distribution of the weighted graph are extracted. The contrast of the node degree and the contrast of the node unit intensity are calculated as the texture feature indexes. The texture feature indexes are used for flotation production state classification and recognition. The experimental results demonstrate that the proposed method can extract froth image texture features accurately and provide effective guidance for flotation production.

QUALITY ASSESSMENT OF BUILDING TEXTURES EXTRACTED FROM OBLIQUE AIRBORNE THERMAL IMAGERY

Abstract. Thermal properties of the building hull became an important topic of the last decade. Combining the thermal data with building models makes it possible to analyze thermal data in a 3D scene. In this paper we combine thermal images with 3D building models by texture mapping. We present a method for texture extraction from oblique airborne thermal infrared images. We put emphasis on quality assessment of these textures and evaluation of their usability for thermal inspections. The quality measures used for assessment are divided to resolution, occlusion and matching quality.

QUALITY ASSESSMENT OF BUILDING TEXTURES EXTRACTED FROM OBLIQUE AIRBORNE THERMAL IMAGERY

Thermal properties of the building hull became an important topic of the last decade. Combining the thermal data with building models makes it possible to analyze thermal data in a 3D scene. In this paper we combine thermal images with 3D building models by texture mapping. We present a method for texture extraction from oblique airborne thermal infrared images. We put emphasis on quality assessment of these textures and evaluation of their usability for thermal inspections. The quality measures used for assessment are divided to resolution, occlusion and matching quality.

Filter Design and Performance Evaluation for Fingerprint Image Segmentation.

Fingerprint recognition plays an important role in many commercial applications and is used by millions of people every day, e.g. for unlocking mobile phones. Fingerprint image segmentation is typically the first processing step of most fingerprint algorithms and it divides an image into foreground, the region of interest, and background. Two types of error can occur during this step which both have a negative impact on the recognition performance: ‘true’ foreground can be labeled as background and features like minutiae can be lost, or conversely ‘true’ background can be misclassified as foreground and spurious features can be introduced. The contribution of this paper is threefold: firstly, we propose a novel factorized directional bandpass (FDB) segmentation method for texture extraction based on the directional Hilbert transform of a Butterworth bandpass (DHBB) filter interwoven with soft-thresholding. Secondly, we provide a manually marked ground truth segmentation for 10560 images as an evaluation benchmark. Thirdly, we conduct a systematic performance comparison between the FDB method and four of the most often cited fingerprint segmentation algorithms showing that the FDB segmentation method clearly outperforms these four widely used methods. The benchmark and the implementation of the FDB method are made publicly available.

Prevalence of severe mitral regurgitation eligible for edge-to-edge mitral valve repair (MitraClip).

We assessed the prevalence of moderately severe or severe mitral regurgitation (MR) justifying edge-to-edge mitral valve (MV) repair (MitraClip(®)) in patients attending the University Hospital Wuerzburg, a tertiary care centre located in Wuerzburg, Germany. Transcatheter edge-to-edge MV repair of advanced MR is a non-surgical treatment option in inoperable and high-risk patients. It is unknown how many patients are potentially eligible for MitraClip(®) since several anatomical prerequisites of the MV apparatus have to be met for optimal treatment results. Using a novel clinical data warehouse we searched for all patients attached to our Department of Internal Medicine from 01/2008 to 01/2012 with moderately severe or severe MR and aged≥18years. The current status of their treatment regime and eligibility for MitraClip(®) was assessed and re-evaluated according to current guidelines and echocardiographic criteria. The search of electronic medical records amongst 43,690 patients employed an innovative validated text extraction method and identified 331 patients with moderately severe or severe MR who had undergone echocardiographic assessment at our institution. Of these, 125 (38%) received MV surgery and 206 (62%) medical therapy only. Most patients not undergoing surgery had secondary MR (70%). After evaluation of medical and echocardiographic data of medically treated patients (n=206), 81 (39%) were potential candidates for MitraClip(®) therapy, and 90 (44%) died during the median follow-up time of 23 months. A large fraction of patients with moderately severe or severe MR but not operated was detected. Medically treated patients had a bad prognosis and about 40% of them were potential candidates for MitraClip(®) therapy.

A Method of Effective Text Extraction for Complex Video Scene

Text information contains important information for video analysis, indexing, and retrieval. Effective and efficient text extraction has been a challenging topic in recent years. Focusing on this issue, a text extraction method for complex video scene is proposed in this paper. Multiframe corner matching and heuristic rules are combined together to detect the text region candidates, which solves the issue of Harris corner filtration for complex video scene and also improves the detection accuracy using multiframe fusion. Local texture description is then used for similarity evaluation judged by SVM. Experimental results for 4 different types of 395-frame video images show the effectiveness of the proposed method compared with 5 existing text extraction methods.

A New Approach to Extract Text from Images based on DWT and K-means Clustering

AbstractText present in image provides important information for automatic annotation, indexing and retrieval. Therefore, its extraction is a well known research area in computer vision. However, variations of text due to differences in orientation, alignment, font, size, low image contrast and complex background make the problem of text extraction extremely challenging. In this paper, we propose a texture-based text extraction method using DWT with K-means clustering. First, the edges are detected from image by using DWT. Then, a small size overlapped sliding window is used to scan high frequency component sub-bands from which texture features of text and non-text regions are extracted. Based on these features, K-means clustering is employed to classify the image into text, simple background and complex background clusters. Finally, voting decision process and area based filtering are used to locate text regions exactly. Experimentation is carried out using public dataset ICDAR 2013 and our own dataset f...

An experiment on wear particle’s texture analysis and identification by using deterministic tourist walk algorithm

Purpose – This study aims to use a deterministic tourist walk to build a system that can identify wear particles. Wear particles provide detailed information about the wear processes taking place between mechanical components. Identification of the type of wear particles by image processing and pattern recognition is key to effective online monitoring algorithm. There are three kinds of particles that are particularly difficult to distinguish: severe sliding wear particles, fatigue spall particles and laminar particles. Design/methodology/approach – In this study, an identification method is tested using the deterministic tourist walking (DTW) method. This study examined whether this algorithm can be used in particle identification. If it does, can it outperform the traditional texture analysis methods such as Discrete wavelet transform or co-occurrence matrix. Different parameters such as walk’s memory size, size of image samples, different inputting vectors and different classifiers were compared. Findings – The DTW algorithm showed promising result compared to traditional texture extraction methods: discrete wavelet transform and co-occurrence matrix. The DTW method offers a higher identification accuracy and a simple feature vector. A conclusion can be drawn that the DTW method is suited for particle identification and can be put into practical use in condition monitoring systems. Originality/value – This paper combined DTW algorithm with wear particle identification problem.

Complex networks-based texture extraction and classification method for mineral flotation froth images

With recent improvements in instrumentation and computer infrastructure, machine vision technologies have produced innovative approaches for controlling and monitoring mineral flotation process. In order to efficiently analyze froth image, froth texture is used to accurately and rapidly extract froth characteristics based on the statistics of image pixels. The characterization and identification of texture require a method that can express the context surrounding each pixel by combining local and global texture characteristics. To extract the distinctive froth texture features in different production states, a novel complex networks-based texture extraction and classification method for froth imaging is proposed in this paper. A network model is constructed by expressing pixels as network nodes and similarities between pixels as network links. This method automatically sets the optimal algorithm parameters for the complex network modeling of the froth images according to bubble sizes by using the Minkowski distance. Energy and entropy measurements are used to quantify the properties of the connectivity and topology of the froth-image network model. Copper froth images at different production states extracted from the flotation monitoring system in a flotation plant are used to test the froth-image network model. The experimental results show that the proposed method accurately describes the froth image texture and also robustly classifies the different production states.

A local binary pattern based texture descriptors for classification of tea leaves

For tea processing production lines, different fresh tea leaves require different processing parameters for the control systems of tea machines. Hence, an effective algorithm for classification of tea leaves will be important for automatic tea processing. However, most of tea classification researches were focused on gross tea, instead of fresh tea leaves. In this paper, a texture extraction method combing a non-overlap window local binary pattern (LBP) and Gray Level Co-Occurrence Matrix (GLCM) has been proposed for green tea leaves classification. By taking advantages of both LBP and GLCM for texture extraction, this method is able to effectively extract texture of tea leaves for classification at low computational cost to meet automatic tea production line requirements. The experiments have been conducted to prove the effectiveness of the proposed method.

Review Paper on Various Methodology of Text Extraction from Image

This review presents the various text extraction techniques and also compares the research results of various researchers in the domain of text extraction. A generic character recognition system has different stages like noise removal, skew detection and correction, segmentation, feature extraction and character recognition. Input is digitized image containing any text, which is preprocessed to segment it into normalized individual word and letters. The OCR, Neural Network, SVM are the various methods for text extraction. Text extraction helps to preserve history by making information efficiently searchable, easily manageable without the need for human labor.

Scene Text Extraction in IHLS Color Space Using Support Vector Machine

Scene text extraction is challenging research area due to variety of image degradations caused by imaging conditions and low cost consumer devices. In this paper we propose text extraction method that uses chroma and lightness component for generation of extraction hypotheses and incorporates SVM (support vector machine) based text detection stage as tool for hypotheses verification. Choice of chroma and lightness components is based on their complementarity with respect to image degradations like shadows and highlights. Another novelty is usage of IHLS color space for text extraction task which is motivated by saturation definition that eliminates instability of this component at low lightness values. Results obtained on the ICDAR 2011 dataset confirm complementarity of chroma and lightness. Compared to the state-of-the-art methods proposed algorithm achieves higher correct recognition rate and comparable total edit distance. DOI: http://dx.doi.org/10.5755/j01.itc.44.1.5757

Recognition of Grocery Products in Images Captured by Cellular Phones

In this paper, we present a robust algorithm to character classes. In addition, our experiments have shown recognize extracted text from grocery product images captured and image processing. HE use of digital cameras to capture text from natural Sciences,3005 Bern, Switzerland (e-mail: farshideh.einsele@bfh.ch). Science, University of Central Florida, Orlando, FL 32816, USA (e-mail: foroosh@cs.ucf.edu). World Academy of Science, Engineering and Technology International Journal of Computer, Information, Systems and Control Engineering Vol:9 No:1, 2015 159 International Scholarly and Scientific Research & Innovation 9(1) 2015 In te rn at io na l S ci en ce I nd ex V ol :9 , N o: 1, 2 01 5 w as et .o rg /P ub lic at io n/ 10 00 02 57 into Fourier domain and ignore the phase information to gain rotation invariant features. Scale and translation invariance is gained with normalization and considering the centroid as the center of the coordinate system. The experimental results show that the centroid and complex coordinate signatures have a high precision and recall rate whereas the curvature and cumulative angular functions deliver unreliable results. Dionisio et al. in [11] also report a contour-based shape classification technique based on polygon approximation that is invariant under rotation and scaling. The vertices of polygon approximation are formed by high curvature points of the profile and are selected by the Fourier transform of the object contour. A series of features are computed from the polygonal approximation and a minimum distance classifier is used for object recognition. Although such contour-based invariants deliver promising results using Fourier descriptors for character recognition and are also reported in the survey of Trier et al. [2], the reported test and training databases are synthetically deformed patterns and the features are invariant with respect to translation, scaling, rotation and do not consider other transformations coming from real world captured images (e.g. shearing, shadowing, bad illumination and perspective distortion). Besides Trier et al.state in [2] that a statistical classification system should consider the so called curse of dimensionality meaning that it should be training-based containing a minimum number of patterns that is 8-10 times bigger than the number of the chosen features. As already stated, when dealing with a database containing real world character images, database generation is an expensive and time-consuming drawback. To sum up, the above mentioned works have been performed either by considering a synthetically degraded database or are training-based approaches. In this paper, we present a method for camera-based character recognition that uses a small real world database extracted from images of grocery products captured by a cellular phone with a resolution of 5 mega pixels. The presented method does not need a training set that should rely on the previously described term of curse of dimensionality. Therefore our proposed method can be applied directly to the extracted text with no use of cost-intensive image enhancement algorithms and delivers promising results. The remainder of this paper is organized as follows: in section II, we introduce the specificities of text in product images. Section III reports about our proposed character recognition algorithm including used feature extraction and classification methods. Section IV presents our evaluation results and section V is about our conclusions and a short sketch of our future works. II. CHALLENGES OF PRODUCT TEXT RECOGNITION We extract text from images taken with cell phones from grocery products. Text extraction from camera-based images is a relatively well researched area with plenty of existing works in the literature [12]. However, text extraction methods from camera-based images is tightly related to a specific application and there does not exist a valid generic method for the extraction of text within different camera-based scenarios. We therefore use a text extraction algorithm that has been developed for the specificities of the text from grocery products and is explained in detail in [13]. The resolution of the used cell phone camera is 5 mega pixels and the images are taken from different angles with the camera having a similar distance to products as the one a common grocery shopper would have when he crosses grocery aisles. The extracted text has mostly a height between 20-50 pixels and characters can be mostly labeled and segmented using connected components algorithms. Table I shows some extracted words in our database.

Integrating multiple texture methods and NDVI to the Random Forest classification algorithm to detect tea and hazelnut plantation areas in northeast Turkey

Hazelnuts and tea are two major agricultural crops grown in the eastern Black Sea region in Turkey. Since this part of Turkey is not industrialized, most of the local people work in agriculture, making hazelnuts and tea a part of their lives. For the government side, it is crucial to keep records of the amount of harvested croplands to implement agricultural policies. In fact, the harvested area and crop type of each cadastral parcel are collected either during cadastral surveys or with the declaration of individual farmers, yet this information is mostly not up-to-date and does not reflect the current land-use status. This study aims to determine the extent and distribution of hazelnuts and tea grown areas using the Random Forest (RF) classification algorithm. Tea and hazelnuts give similar spectral reflectance values to surrounding vegetation, which makes it difficult to distinguish them using only their spectral properties. To tackle this problem, the normalized difference vegetation index (NDVI) and texture extraction methods such as the Grey Level Co-occurrence Matrix (GLCM) and Gabor filter were integrated with the RF algorithm, and their contributions to the success of the RF classification method were examined. WorldView-2 satellite images, which have eight multispectral bands (MS: 2 m) and one higher spatial resolution panchromatic band (PAN: 0.5 m), were used. Since the study area contains agricultural products grown in different seasons, satellite images belonging to both summer and winter periods were used. Preliminary results acquired using only spectral values indicated that the RF method gives 79.05% and 71.84% overall accuracies for summer and winter periods, respectively. Integrating texture information improves the performance of the RF algorithm such that the overall classification accuracies are increased to 83.54% and 87.89% when texture information extracted with GLCM and the Gabor filter is added. The classification performance of the winter image is also boosted to be 77.41% and 79.73% with the contribution of texture information obtained with GLCM and the Gabor filter, respectively. Finally, produced thematic maps were compared with the latest cadastral maps to validate classification results with ground truth data. The obtained results reveal the success of integrating texture features in classification since the created thematic maps are consistent with actual land use. The results also show that the crops grown on some cadastral parcels are not coherent with the most current cadastral database, which implies that the cadastral maps need to be updated.