Text Size Research Articles

Information for Contributors

Information for Contributors

Arbitrarily Shaped Scene Text Detection with a Mask Tightness Text Detector.

Scene text in the environment is complicated. It can exist in arbitrary text fonts, sizes or shapes. Although scene text detection has witnessed considerable progress in recent years, the detection of text with complex shapes, especially curved text, remains challenging. Datasets with adequate samples to overcome the problem presented by curved text (or other irregularly shaped text) have been introduced only recently; however, the performance of the reported methods on these datasets is unsatisfactory. Therefore, detecting arbitrarily shaped text remains a challenging. This motivated us to propose the Mask Tightness Text Detector (Mask TTD) to improve text detection performance. Mask TTD uses a tightness prior and text frontier learning to enhance pixel-wise mask prediction. In addition, it achieves mutual promotion by integrating a branch for the polygonal boundary of each text region, which significantly improves the detection performance of arbitrarily shaped text. Experiments demonstrate that Mask TTD can achieve state-ofthe-art performance on existing curved text datasets (CTW1500, Total-text, and CUTE80) and three common benchmark datasets (RCTW-17, MSRA-TD500, and ICDAR 2015). It is worth mentioning that on CTW1500, our method can outperform previous methods, especially at higher intersection over union (IoU) thresholds (16% higher than the next-best method with an IoU threshold of 0.8), which demonstrates its potential for tight text detection. Moreover, on the largest Chinese-based dataset RCTW-17, Mask TTD outperforms other methods by a large margin in terms of both the Average Precision and F-measure, showing its powerful generalization ability.

Natural scene text localization using edge color signature

Localizing text regions in images taken from natural scenes is one of the challenging problems dueto variations in font, size, color and orientation of text. In this paper, we introduce a new concept socalled Edge Color Signature for localizing text regions in an image. This method is able to localizeboth Farsi and English texts. In the proposed method rst a pyramid using different scales of theinput image is created. Then for each level of the pyramid an edge map is extracted. Afterward,several geometric features are employed to lter out the non-text edges from the extracted edges.At this stage we describe an edge using colors of its neighboring pixels. We use the mean-Shiftalgorithm to obtain the color modes surrounding each edge pixel. Subsequently, the connectededge pixels with similar color signatures are clustered using Single-Linkage clustering algorithm toconstruct meaningful groups. Finally, each of the clusters is labeled as text or non-text using anMLP based cascade classi er. The proposed method has been evaluated on well-known ICDAR 2013and our Farsi dataset, the result is very promising.

An algorithm to identify text marked in rock core pictures with machine learning algorithm

The aim of our project is to extract numerical descriptors or gather feature of rock samples from their images made under microscope with computer vision algorithm which can overcome the bottleneck of traditional method. It is impossible to solve it without getting the imformation of microscope magnification of images when they were made. That is why this paper is about an algorithm to identify the text used to mark microscope magnification information in rock core sample pictures. This is a sub project. As locations of text in these pictures are not fixed, font of text is not fixed, font size of text is not fixed, a method of two steps is designed to identify the text from the background. It works as follows: first, find the location of text; second, recognize the text with Convolution Neural Network. Experimental results demonstrate the potential value of the proposed approach to identify the magnification information in rock core sample pictures and to apply as a funtion to a program of our future project.

Performance Analysis of Data Encryption Algorithm

In this paper, author have performed experimental analysis of Jumbling-Salting (JS) algorithm for larger text size. In the earlier research work, JS Algorithm was symmetric-password encryption algorithm. JS algorithm consists of two prominent cryptographic processes namely Jumbling and Salting. Jumbling consists of three major randomized processes viz. Addition, Selection and Reverse. Jumbling process jumbles the random characters into password string. Salting process adds a random string based on some timestamp value. The output of Jumbling and Salting process is given to predefined AES block to perform 128-bit key encryption to maintain the cipher text size uniform. In this research, the capability of JS algorithm is enhanced. The paper therefore shows the performance of JS algorithm regarding length of cipher text size with respect to AES and DES algorithms. This extended research ensures that JS algorithm is not only suitable for smaller text like password, pin, passcode etc. but it is also favorable for larger texts.

YAKE! Keyword extraction from single documents using multiple local features

As the amount of generated information grows, reading and summarizing texts of large collections turns into a challenging task. Many documents do not come with descriptive terms, thus requiring humans to generate keywords on-the-fly. The need to automate this kind of task demands the development of keyword extraction systems with the ability to automatically identify keywords within the text. One approach is to resort to machine-learning algorithms. These, however, depend on large annotated text corpora, which are not always available. An alternative solution is to consider an unsupervised approach. In this article, we describe YAKE!, a light-weight unsupervised automatic keyword extraction method which rests on statistical text features extracted from single documents to select the most relevant keywords of a text. Our system does not need to be trained on a particular set of documents, nor does it depend on dictionaries, external corpora, text size, language, or domain. To demonstrate the merits and significance of YAKE!, we compare it against ten state-of-the-art unsupervised approaches and one supervised method. Experimental results carried out on top of twenty datasets show that YAKE! significantly outperforms other unsupervised methods on texts of different sizes, languages, and domains.

High performance parallel KMP algorithm on a heterogeneous architecture

String matching algorithm is widely used in many application areas such as bio-informatics, network intrusion detection, computer virus scan, among many others. KMP (Knuth-Morris-Pratt) algorithm is commonly used for its fast execution time compared with many other string matching algorithms when applied to large input texts. However, the performance of the KMP algorithm is limited when the input text size increases significantly beyond a certain limit. In this paper, we propose a high performance parallel KMP algorithm on the Heterogeneous High Performance Computing (HPC) architecture based on the general-purpose multicore microprocessor and the Graphic Processing Unit (GPU) using OpenCL as the GPGPU (General Purpose computing using Graphic Processing Unit) platform. The proposed parallel KMP algorithm mainly focuses on optimizing the CPU-GPU memory hierarchy by overlapping the data transfer between the CPU memory and the GPU memory with the string matching operations on the GPU. It also optimizes the allocation of the work-groups and the work-items, and places the pattern data and the Failure Table to the on-chip shared-memory of the GPU. The experimental results show that the optimized parallel KMP algorithm leads up to ~8 times faster execution time.

Extractive Text Summarization using Deep Natural Language Fuzzy Processing

Text summarization is most trending research areas in a modern context. The main aim of this project is to reduce text size while preserving the information underlying into it. In summary construction level, in general, given complex task which are basically will involve with deep natural language fuzzy processing methodologies. In general, an extractive based summary method is the very simple original text of subset of which will not guarantee as best narrative coherence output, because they are most conveniently representing an approximate summarized content from given text-based only on relevance judgment. In an automatic process of fuzzy summarization which is divided into the following steps: Pre-processing (sentence segmentation, tokenization, stop words removal), Feature Extraction, Sentence Scoring, Sentence Ranking and Summary Extraction.

Method for Detecting Chinese Texts in Natural Scenes Based on Improved Faster R-CNN

In this study, a natural scene text detection method based on the improved faster region-based convolutional neural network (R-CNN) is proposed. This method extracts image features with the Inception-ResNet architecture, adopts a region proposal network to generate region proposals for the extracted features, merges the fine-tuned features with the region proposals, and finally, uses Fast R-CNN to classify and locate text. The proposed method solves the problems of varying text sizes and the text being obscured in the image. Compared with the original Faster R-CNN, the multilevel Inception-ResNet network model presented in this study can extract deeper text features. The extracted feature map is further sparsely represented by Reduction B, Inception ResNet C and Avg Pool, and then is fused with text regions obtained by the text feature mapping lower layer network to acquire the exact text regions. The text detection method presented in this study is tested on the 2017 dataset of ICDAR2017 Competition on Reading Chinese Text in the Wild (RCTW-17), which contains a large number of distorted, blurry, different scale and size texts. An accuracy of 76.4% is achieved in this platform, thereby proving the efficiency of the proposed method.

Experimental and theoretical study of corrosion inhibition performance of N-phenylthiourea for mild steel in hydrochloric acid and sodium chloride solution.

The corrosion inhibition performance of N-phenylthiourea for mild steel in solutions of 1.0M hydrochloric and 3.5% wt. sodium chloride was investigated using tandem potentiodynamic polarization curves, scanning electron microscope analysis and quantum chemical calculations. The inhibition efficiency values of N-phenylthiourea in acidic solution and in salt solution were 94.95% and 55.70%, respectively. The results show that the corrosion inhibitor ability of N-phenylthiourea was better than that of urotropine under the same conditions. The adsorption of N-phenylthiourea on the mild steel electrode surface obeys the Langmuir adsorption isotherm in acidic solution and modified Langmuir adsorption isotherm in salt solution. The results of quantum chemical calculations and molecular dynamics simulations were found to be in accord with experimental data. Graphical abstract Corrosion inhibition performance of N-phenylthiourea for mild steel in hydrochloric acid and sodium chloride solution Figure A contains poor quality and small text inside the artwork. Please do not re-use the file that we have rejected or attempt to increase its resolution and re-save. It is originally poor, therefore, increasing the resolution will not solve the quality problem. We suggest that you provide us the original format. We prefer replacement figures containing vector/editable objects rather than embedded images. Preferred file formats are eps, ai, tiff and pdf.Figure A now has been prepared. The font size of text is increased except for the SEM image.

Implementation Challenges in Information Retrieval System

Natural scenes text recognition has become increasingly attractive in recent years. Several research results have been developed on different languages for the sake of text recognition from natural scenes. However, there is no system for identifying Urdu text from natural scenes. This paper presents a framework steps involved in recognition of Urdu text from landscape images. Furthermore, the paper focus on issues and challenges faced during experimentation and implementation. Challenges includes multi-script text on a single image, different text font with variation in text sizes, light shadow, reflection, text color, text contrast with background, words or ligature missing, blur, low resolution , skew and neon signboard. In the end, we also give a summary of issues and challenges of Urdu Text retrieval system from natural scenes.

Aligning content rendering resolution and feature size with display capability in near‐eye display systems

AbstractIn order to create a satisfying experience with near‐eye displays, the content must be adapted to be legible on the display used. New displays are using subpixel arrangements that can limit the minimum resolvable feature size to something greater than with the conventional RGB stripe arrangement. We conducted an experiment to measure the minimum and preferred size of text in two virtual reality (VR) displays systems and find that the text size is display limited. We then measure several displays with different pixel arrangements to determine whether the subpixel arrangement could impact legibility. We propose several Fourier metrics that can be computed from the measured data to categorize the capability of the display and describe a framework for selecting the appropriate content from a set of discrete tiers.

Study of Optimal Text Size Phenomenon in Zipf–Mandelbrot’s Distribution on the Bases of Full and Distorted Texts. Author’s Frequency Characteristics and derivation of Hapax Legomena

ABSTRACTThis paper explores word-frequency patterns when considering text length, authorship, and random distortion of texts. Through a series of experiments, we determined an optimal text size, a phenomenon that was predicted by George Zipf, which sees a minimal discrepancy between calculated and observed frequencies. A graphic representation allowed a plausible explanation behind the existence of this phenomenon. Working on the assumption that distorted texts might disobey Zipf’s Law, we explored correlations among frequencies and text entirety compared with text distortions. Results reveal the crucial role of text length for maintaining Zipfian distribution: randomly chosen sets of words and fragmentary texts of optimal size still obey Zipf’s Law. Findings show that authorship manifests itself through the author constant, defined as the relative frequency of the most frequent words, which remains constant throughout the works of any given author, including randomly chosen text chunks and fragments of sentences of various sizes.

Perspectives on Acceptance and Use of a Mobile Health Intervention for the Prevention of Atherosclerotic Cardiovascular Disease in Singapore: Mixed-Methods Study.

BackgroundCardiovascular disease, including atherosclerotic cardiovascular disease (ASCVD), is a growing public health threat globally and many individuals remain undiagnosed, untreated, and uncontrolled. Simultaneously, mobile health (mHealth) interventions using short messaging service (SMS) have gained popularity globally. There is an opportunity for innovative approaches such as mHealth to encourage and enable adherence to medications for ASCVD and its risk factors.ObjectiveThis study aimed to understand mobile technology acceptance, use, and facilitating conditions among the study population ahead of the design of an mHealth intervention.MethodsUsing data from a mixed-methods study conducted in Singapore, we conducted a cross-sectional survey with 100 participants and in-depth, semistructured interviews with 20 patients. All participants were over the age of 40 years with ASCVD or its risk factors. Interviews were conducted in English and Mandarin and if needed translated to English. Nvivo 11 (QSR International) was used for analyses.ResultsParticipants reported their perspectives on technology use and preferences, including low or sporadic mobile phone use and usability concerns including small screen and text size, among others; the benefit of previous mHealth use in creating a favorable opinion of SMS for health information; trust in both the source of mHealth SMS, as well as in treatment; the formation of habits; and fear of sequelae or death for facilitating intention to use an mHealth intervention and adhere to medication. We also highlighted a case that underscored the importance of the period after diagnosis in habit forming as an opportunity for an mHealth intervention.ConclusionsWe explored both technology- and adherence-related factors that influence a patient’s intention to use an mHealth intervention for adherence to ASCVD medication in Singapore. We highlighted the importance of identifying the right opportunity to engage with patients and promote an mHealth intervention for adherence, such as immediately following diagnosis when patients are establishing medication-taking habits.

Can dependency distance and direction be used to differentiate translational language from native language?

Translational language is called “the third code” or “the third language” by some researchers. However, whether “the third code” truly exists or not has been an issue of controversy for decades ever since it was proposed. In this study, in the framework of dependency grammar, we use mean dependency distances (MDD) and dependency direction for analyzing translated texts and native texts in an attempt to differentiate translational language from native language in a quantitative way, while displaying holistic features. We built a dependency-annotated treebank consisting of translated English texts, native English texts and native Chinese texts. To eliminate the influence of sentence length and text size on the MDD and dependency direction, we calculated the MDD of each sentence length, and made sure each text was similar in size. We found that: (1) the MDD of the translated English texts was significantly different from that of native English texts, with the MDD of the former much longer than that of the latter; (2) texts of the two types were also significantly different in dependency direction, with the translated English texts being more head-initial compared with the native English texts. Differences in the MDD is probably influenced by the bigger cognitive effort in dealing with Chinese source language and difference in dependency direction is probably due to translators’ tendency of putting the informational focus ahead while making decisions. These findings suggest that (1) dependency distance and direction can be used to differentiate translational language from native language; (2) typologically speaking, translational language, a third language which is distinctive from native language may exist; (3) translational language is produced under the combined effect of cognitive effort and decision-making process by translators; and 4) syntactic quantitative methods can be applied to translation studies.

Performance analysis of hybrid cryptosystem - A technique for better security using blowfish and RSA

Security is the most concerned topic in this today’s world of Information and Communication Technology. Single layer security can be breached easily so hybrid cryptographic system has been introduced. Hybrid system can be made by combining two or more algorithms of similar or dissimilar types. Here, combing of Blowfish algorithm along with RSA algorithm is performed. After combination their performance analysis is done based on five parameters namely Throughput, Encryption time, Decryption time, Total Execution time and Plain text size to cipher text size ratio with different key size of Blowfish algorithm ranging from 32 bit-448 bit. Among these 5 different hybrid cryptosystem we found that the Blowfish RSA system with key size 448 bit has better performance than all other bit size.

Convolutional Neural Network Based Text Steganalysis

The prevailing text steganalysis methods detect steganographic communication by extracting hand-crafted features and classifying them using SVM. However, these features are designed based on the statistical changes caused by steganography, thus they are difficult to adapt to different kinds of embedding algorithms and the detection performance is heavily dependent on the text size. In this letter, we propose a novel text steganalysis model based on convolutional neural network, which is able to capture complex dependencies and learn feature representations automatically from the texts. First, we use a word embedding layer to extract the semantic and syntax feature of words. Second, the rectangular convolution kernels with different sizes are used to learn the sentence features. To further improve the performance, we present a decision strategy for detecting the long texts. Experimental results show that the proposed method can effectively detect different kinds of text steganographic algorithms and achieve comparable or superior performance for a wide variety of text sizes compared with the previous methods.

Visual requirement for Chinese reading with normal vision.

ObjectivesReading performance has been considered as an effective functional endpoint for low vision. Contrary to many extensive studies for reading performance in English, there are few systematic studies for Chinese reading.MethodsIn the present study, the reading performance of 30 normally sighted Chinese college students was systematically investigated. All participants passed the equivalent test of Cambridge ESOL PET in China. The reading speeds for Chinese and English text at a variety of text sizes were measured with rapid serial visual presentation (RSVP). The threshold acuities for Chinese characters and English letters were measured. Maximum reading speed, critical font size, and critical acuity reserve were derived according to the individual's reading speed curve.ResultsThe maximum reading speed for Chinese characters was 259.5 ± 38.2 characters/min, which was significantly faster than that for English letters (135.7 ± 18.5 words/min, p = 2.8 × 10‐18). The critical font size for Chinese characters was larger than that for English letters (24.2 ± 2.8 arcmin vs. 20.7 ± 1.0 arcmin, p = 1.6 × 10‐7). Interestingly, the critical acuity reserve was similar for these two languages (3.4 ± 0.4 for Chinese and 3.4 ± 0.2 for English, p = 0.4).ConclusionThe present study provides the first step for establishing visual functional endpoints for Chinese reading. Our findings pose rigorous constrains on present theories in language information processing and brain plasticity.

Active photonic sensing for super-resolved reading performance in simulated prosthetic vision.

In this work, we study the enhancement of simulated prosthetic reading performance through "active photonic sensing" in normally sighted subjects. Three sensing paradigms were implemented: active sensing, in which the subject actively scanned the presented words using the computer mouse, with an option to control text size; passive scanning produced by software-initiated horizontal movements of words; and no scanning. Our findings reveal a 30% increase in word recognition rate with active scanning as compared to no or passive scanning and up to 14-fold increase with zooming. These results highlight the importance of a patient interactive interface and shed light on techniques that can greatly enhance prosthetic vision quality.

PVOPM: Verifiable Privacy-preserving Pattern Matching with Efficient Outsourcing in the Malicious Setting

Outsourced pattern matching delegates the task of finding all positions pattern $P$ P appears in text $T$ T to the cloud from resource-constrained devices. Unfortunately, it has brought a series of security and privacy issues. Most of the state-of-the-art either disclosed text/pattern privacy or exploited the computationally-intensive techniques of public key fully homomorphic encryption (FHE), commitment schemes and zero knowledge proof to achieve both text/pattern privacy and verifiability. To address these issues, as a building block, an efficient privacy preserving verifiable outsourced discrete fourier transform protocol OVFT is first devised based on any one-way trapdoor permutation (OWTP). Based on OVFT, we propose an efficient secure verifiable outsourced polynomial multiplication protocol OPVML which is further exploited in designing our final protocol PVOPM for verifiable privacy-preserving outsourced pattern matching. Without exploiting public key FHE, the proposed PVOPM achieves both verifiability and text/pattern privacy against the collusion between the cloud and malicious receiver/sender, by generating authentication proofs of constant size and executing constant times of any one-way trapdoor permutation, independent to both the text size $n$ n and the pattern size $m$ m . Finally, formal security proof under universal composable (UC) model and extensive evaluations demonstrate the efficiency and practicability of our proposed PVOPM.