Text Localization Research Articles

A Study of Game Localization Translation under the Perspective of Skopos Theory--Taking the Game Red Dead Redemption 2 as an Example

As economic globalization continues to advance, the international market for video games is rapidly growing. For game companies aiming to expand their business globally, the localization of game translations is a crucial aspect. A well-executed localization can greatly boost a game’s international sales and appeal, attracting players from diverse regions. Thus, game localization is a vital strategy for companies seeking success on the global stage. Skopos theory, a key framework in translation studies, offers valuable insights for different types of translation work. This paper applies the principles of Skopos theory to examine the Chinese localization of various texts in the large-scale single-player game Red Dead Redemption 2. The analysis reveals that, under the guidance of Skopos theory, the localization enhances both the gaming experience and the sense of immersion for Chinese players. During the localization process, the translation, adaptation, or even rewriting of the original text is performed with the goal of aligning the final product with the cultural and linguistic preferences of local players.

A novel domain independent scene text localizer

Text localization across multiple domains is crucial for applications like autonomous driving and tracking marathon runners. This work introduces DIPCYT, a novel model that utilizes Domain Independent Partial Convolution and a Yolov5-based Transformer for text localization in scene images from various domains, including natural scenes, underwater, and drone images. Each domain presents unique challenges: underwater images suffer from poor quality and degradation, drone images suffer from tiny text and loss of shapes, and scene images suffer from arbitrarily oriented, shaped text. Additionally, license plates in drone images may not provide rich semantic information compared to other text types due to loss of contextual information between characters. To tackle these challenges, DIPCYT employs new partial convolution layers within Yolov5 and integrates Transformer detection heads with a novel Fourier Positional Convolutional Block Attention Module (FPCBAM). This approach leverages common text properties across domains, such as contextual (global) and spatial (local) relationships. Experimental results demonstrate that DIPCYT outperforms existing methods, achieving F-scores of 0.90, 0.90, 0.77, 0.85, 0.85, and 0.88 on Total-Text, ICDAR 2015, ICDAR 2019 MLT, CTW1500, Drone, and Underwater datasets, respectively.

An End-to-End Scene Text Recognition for Bilingual Text

Text localization and recognition from natural scene images has gained a lot of attention recently due to its crucial role in various applications, such as autonomous driving and intelligent navigation. However, two significant gaps exist in this area: (1) prior research has primarily focused on recognizing English text, whereas Arabic text has been underrepresented, and (2) most prior research has adopted separate approaches for scene text localization and recognition, as opposed to one integrated framework. To address these gaps, we propose a novel bilingual end-to-end approach that localizes and recognizes both Arabic and English text within a single natural scene image. Specifically, our approach utilizes pre-trained CNN models (ResNet and EfficientNetV2) with kernel representation for localization text and RNN models (LSTM and BiLSTM) with an attention mechanism for text recognition. In addition, the AraElectra Arabic language model was incorporated to enhance Arabic text recognition. Experimental results on the EvArest, ICDAR2017, and ICDAR2019 datasets demonstrated that our model not only achieves superior performance in recognizing horizontally oriented text but also in recognizing multi-oriented and curved Arabic and English text in natural scene images.

English foreign language learners’ perceptions of cultural elements in Chinese EFL textbooks

Our study aims to culturally analyze the textbook series in the context of China using a questionnaire taken from Kilickaya (2004). To this end, 635 EFL students participated and filled out a questionnaire. In addition, 31 students were randomly selected for a semi-structured interview. The results of the study indicated that students’ perceptions of the meaning and purpose of textbooks were found to be positive. As for materials on culture, it is argued that textbooks should include both macro and micro-cultures. The findings showed that the teaching contributions did a good job of presenting various perceptions from the students' points of view. Students felt that the textbooks represented diverse societies and were relevant to their home culture, but they felt that the manuals were less effective in presenting the values of social groups. The results demonstrated that the same location of images and text in textbooks is usually the display index. EFL textbooks for language learners in China are mainly composed of index relationships between texts, pictures, and tasks, showing a focus on descriptive meaning. Ultimately, the outcomes of this inquiry provide practical implications for EFL stakeholders.

TRANSLATING ANIME AND ITS DIFFICULTIES

This study examines various translation strategies employed by English and Russian translators, with a focus on their application in the localization and adaptation of multi-semiotic video-verbal texts, such as animated films. Drawing from the classifications of translation strategies by English linguist and translator P. Newmark, as well as insights from translation theorist I. S. Alekseeva's research explores eight key strategies.

References to Scripture in the Old Polish History of the Three Kings. Quotations, Allusions, Echoes

The article analyzes the references to Scripture in Historyja trzech kroli (History of the Three Kings), an Old Polish apocryphal work preserved in a copy from 1544. The references to the Bible are divided into quotations, allusions and echoes. Quotations were introduced either by reference to a particular location or author of the biblical text, or to an unspecified “prophet” or “evangelist.” Sometimes they were not signaled in any way. Then the quotation could only be identified by recognizing the passage to which it refers. Allusions to the Old and New Testaments mostly consisted of references to characters, places, and events, without quoting the text of Scripture itself. The featured echoes most often consisted of narrative parallelisms between the story of the Wise Men from the East and the first Church. References to Scripture in the apocryphal work served to emphasize typology and create parallelisms between biblical history and the story of the three kings. The text emphasizes the role of the Gentiles in salvation history while belittling the role of Israelites.

Фиты в дореформенных и пореформенных списках нотированного Октоиха XVII в.

The article is devoted to a significant element of znamenny chant - fitnoy singing. It examines both the lettering and chanting of fitts in the notated Octoechos, in particular, how the general patterns of fitt singing are manifested in the Octoechos. This work shows that the use of fit in the chants of the Octoechos does not depend directly on the genre of the chant, although certain regularities can be traced. In addition, the number of fits is distributed unevenly among the voices, just as in other books and in theoretical manuals - fitniki. At the same time, some fitas are repeated more than once within a vowel section. Comparison of lists close in time and type allows us to identify "author's" variants of the arrangement of fittas or chants of fittas, although in general variants of chants for fittas in the margins or in the banner line of Octoechos chants are rarely recorded. Special attention is paid to the degree of preservation of the chants in connection with the process of correction of liturgical books in the middle of the 17th century, which also affected notated manuscripts. In this connection the lists of the pre-reform, post-reform and transitional editions of the Octoechos are compared diachronically. The comparison shows that most of the fictional formulae were preserved in the process of correcting the notated Octoechos, and, as a rule, their location in the poetic text did not change either. These conclusions testify to the preservation of the tradition, which does not exclude the creative approach of the masters in some cases.

ViTEraser: Harnessing the Power of Vision Transformers for Scene Text Removal with SegMIM Pretraining

Scene text removal (STR) aims at replacing text strokes in natural scenes with visually coherent backgrounds. Recent STR approaches rely on iterative refinements or explicit text masks, resulting in high complexity and sensitivity to the accuracy of text localization. Moreover, most existing STR methods adopt convolutional architectures while the potential of vision Transformers (ViTs) remains largely unexplored. In this paper, we propose a simple-yet-effective ViT-based text eraser, dubbed ViTEraser. Following a concise encoder-decoder framework, ViTEraser can easily incorporate various ViTs to enhance long-range modeling. Specifically, the encoder hierarchically maps the input image into the hidden space through ViT blocks and patch embedding layers, while the decoder gradually upsamples the hidden features to the text-erased image with ViT blocks and patch splitting layers. As ViTEraser implicitly integrates text localization and inpainting, we propose a novel end-to-end pretraining method, termed SegMIM, which focuses the encoder and decoder on the text box segmentation and masked image modeling tasks, respectively. Experimental results demonstrate that ViTEraser with SegMIM achieves state-of-the-art performance on STR by a substantial margin and exhibits strong generalization ability when extended to other tasks, e.g., tampered scene text detection. Furthermore, we comprehensively explore the architecture, pretraining, and scalability of the ViT-based encoder-decoder for STR, which provides deep insights into the application of ViT to the STR field. Code is available at https://github.com/shannanyinxiang/ViTEraser.

LRANet: Towards Accurate and Efficient Scene Text Detection with Low-Rank Approximation Network

Recently, regression-based methods, which predict parameterized text shapes for text localization, have gained popularity in scene text detection. However, the existing parameterized text shape methods still have limitations in modeling arbitrary-shaped texts due to ignoring the utilization of text-specific shape information. Moreover, the time consumption of the entire pipeline has been largely overlooked, leading to a suboptimal overall inference speed. To address these issues, we first propose a novel parameterized text shape method based on low-rank approximation. Unlike other shape representation methods that employ data-irrelevant parameterization, our approach utilizes singular value decomposition and reconstructs the text shape using a few eigenvectors learned from labeled text contours. By exploring the shape correlation among different text contours, our method achieves consistency, compactness, simplicity, and robustness in shape representation. Next, we propose a dual assignment scheme for speed acceleration. It adopts a sparse assignment branch to accelerate the inference speed, and meanwhile, provides ample supervised signals for training through a dense assignment branch. Building upon these designs, we implement an accurate and efficient arbitrary-shaped text detector named LRANet. Extensive experiments are conducted on several challenging benchmarks, demonstrating the superior accuracy and efficiency of LRANet compared to state-of-the-art methods. Code is available at: https://github.com/ychensu/LRANet.git

Cartoon single-image super-resolution approach based on generative adversarial network

In recent years, the study of a single image super-resolution (SISR) is crucial to improving image resolution and using hardware technology to improve image resolution. SISR is widely used in satellite remote sensing, video surveillance, and medical image processing because it mainly relies on deep learning algorithms to realize the conversion from low-resolution (LR) images to high-resolution images. It has the advantages of low cost, simple operation, and high efficiency. This paper proposes an image super-resolution method based on a generative adversarial network named text localization generative adversarial nets (TLGAN) model. The method is improved based on super-resolution generative adversarial networks (SRGAN), and the batch normalization layer is removed, which significantly reduces the computational burden of the model. In TLGAN model, we used the transfer learning method to pre-trained the model on the large dataset ImageNet, and then apply the pre-trained model to the cartoon image data set animes to achieve image super-resolution. Experimental results report that the proposed method has the advantages of fast running speed and excellent visual perception of super-resolution images compared with bicubic interpolation and SRGAN method.

The preferences of people in Australia to respond and engage with advertisements to promote reproductive health: Results of a discrete choice experiment

ObjectivesThe health of people prior to pregnancy impacts pregnancy outcomes and childhood health, making the preconception period an important time to optimise health behaviours. Low awareness of the importance of this issue is a recognised barrier to achieving good preconception health. Public health messaging can help to address this barrier. MethodsA discrete choice experiment to assess the preferences of people of reproductive age for a health promotion advertisement for preconception health was conducted. Attributes of the advertisement image, title, additional text content and positioning, and the location of advertisement were assessed by fitting a mixed logit model to the choices made. ResultsThree hundred and thirty-four responses were obtained, from people of reproductive age, both planning and not planning a pregnancy, in Australia. Participants placed most importance on the image, and the location in which they saw the advertisement. An image of adult and baby hands was preferred to adult hands only, and healthcare settings were preferred to more general media locations such as advertising online or on public transport. Preference was also given to the advertisement title of “Healthy you, Healthy baby”, closely followed by “Are you ready for pregnancy?”. The location and content of additional text did not significantly impact engagement with the advertisement. ConclusionThe image and title on the advertisement, and the locations in which they are placed were the most significant features to impact engagement with a health promotion advertisement for preconception health. This can inform health promotion efforts for preconception health.

Tracking e-reading behavior: uncovering the effects of task context, electronic experience, and motivation

Although electronic reading of fiction has become mainstream, little is known about how electronic texts are read, and whether this behavior is connected to readers’ motivation or experience reading electronically. To address this gap, sixty undergraduate students’ reading behavior was tracked while reading a 15-page short story. A novel method was used to study participants’ frequency of task-switching, their reading speed, and navigational patterns unobtrusively, outside of the lab. Reading behavior was analyzed by two multilevel models to assess (1) whether variance in behavior could be predicted by the task context, such as location in text or timing of reading sessions, and (2) whether behavior was connected to participants’ situational motivation to read the short story, their contextual motivation toward reading as an activity, or their task-relevant electronic experience. Our results showed that highly experienced and avid readers reacted to text difficulty more adaptively, indicating that motivation and electronic experience may have a key role in supporting comprehension. In contrast, situational motivation was not associated with reading behavior, contrary to our expectations. These findings provide a significant contribution to our understanding of e-reading, which can be used as a foundation to support recreational reading engagement on digital devices.

Text Localization and Enhancement of Mobile Camera based Complex Natural Bilingual Text Scene Images

In this paper, we proposed a text localization and enhancement of the Mobile camera based complex natural scene images using fixed coordinates for removal of unwanted non text regions and making it localized. The morphological erosion technique is applied for the smoothness of foreground text regions. Here, the experimental working nature of the proposed method is first localized color input image is converted into a gray level, then a median filter is applied for the removal of noises. After that, the input image is further converted into binary image. In addition to this, morphological dilation and erosion operations also applied for connecting the broken pixels. Further, the connected component segmentation technique is applied to extract the foreground text regions from the binary input image. The real time input samples are collected by using mobile camera from the various categories of natural scenes of monolingual/bilingual text such as Stone, Wall, Poster and Iron-based. A total of 140 real time samples of natural scene text images are considered for the experimental setup. The experimental results shown that 132 real time sample images are localized and then obtained the accuracy of 94.24%. Further, bilingual text regions are enhanced using hybrid approach and obtain the segmentation accuracy of 83.17%. The novelty of the paper is working on the real time data set and it contains a mixture of printed and handwritten natural scene text images.

Towards AI-Enabled Approach for Urdu Text Recognition: A Legacy for Urdu Image Apprehension

Recognizing Urdu text in natural images is more challenging as compared to other languages, such as English, due to the cursive nature of Urdu script. However, Urdu scene text has not received enough attention from both industry and academia due to the lack of the dataset of Urdu text. We propose a large-scale Urdu Scene Text Dataset (USTD) to address this problem, which is designed for Urdu scene text detection and recognition. The proposed dataset contains 29674 text annotations (17877 Urdu and 11797 English), 749725 characters in 6389 images. It covers a wide variety of text images with both Nastaleeq and Naskh writing styles, taken from different streets and roads of Pakistan. The vast diversity of this dataset makes it a benchmark to work on and train robust neural networks for the detection and recognition of cursive text. Besides, baseline results are also provided with several state-of-the-art networks, including TextBoxes++, Seglink, DB(ResNet-50) and EAST for text localization and Convolutional Recurrent Neural Network (CRNN) for text recognition. To further evaluate the performance of these models, we have used the most popular evaluation matrices of precision, recall, and F-measure. Our experimental outputs reveal that an end-to-end combination of DB(ResNet-50) and CRNN provides the best results with precision, recall, and F-measure of 0.7526, 0.5974, and 0.6660, respectively.

Classification and Localization of Arabic Handwritten Text in KHATT Dataset Based on Faster R-CNN

Arabic handwriting recognition is an important research area in computer vision. Due to the complexity of the Arabic script, this is an arduous task. Several approaches have been proposed to address this challenge, including deep learning algorithms. Despite their efficiency, these approaches present some limitations such as the use of lexicon-driven models, the need of a lot of data for training and the huge computational cost. We propose, in this paper, two novel models based on the robust Faster Region-Convolution Neural Network (Faster R-CNN) to recognize Arabic handwritten sentences. The Faster R-CNN is commonly used to detect objects in images and depends on region proposal algorithms. These models use the pre-trained VGG 16 and ResNet50. Moreover, they use a soft non-maximum suppression strategy (Soft-NMS) in the post-processing stage instead of the traditional NMS to increase the detection rate of overlapping items. The models will be trained independently to identify words in the text lines and tested using sentences from the (KFUPM Handwritten Arabic TexT) KHATT dataset. It will be shown that the faster R-CNN models allow greater accuracy and effectiveness in handwriting recognition. They achieve accuracy rates of 99% and 98% for the two networks, VGG16 and ResNet50, respectively

Natural scene text localization and detection using MSER and its variants: a comprehensive survey

Natural scene text localization and detection using MSER and its variants: a comprehensive survey

An Automatic Privacy-Aware Framework for Text Data in Online Social Network Based on a Multi-Deep Learning Model

With the increasing severity of user privacy leaks in online social networks (OSNs), existing privacy protection technologies have difficulty meeting the diverse privacy protection needs of users. Therefore, privacy-aware (PA) for the text data that users post on OSNs has become a current research focus. However, most existing PA algorithms for OSN users only provide the types of privacy disclosures rather than the specific locations of disclosures. Furthermore, although named entity recognition (NER) technology can extract specific locations of privacy text, it has poor recognition performance for nested and interest privacy. To address these issues, this paper proposes a PA framework based on the extraction of OSN privacy information content. The framework can automatically perceive the privacy information shared by users in OSNs and accurately locate which parts of the text are leaking sensitive information. Firstly, we combine the roformerBERT model, BI_LSTM model, and global_pointer algorithm to construct a direct privacy entity recognition (DPER) model for solving the specific privacy location recognition and entity nesting problems. Secondly, we use the roformerBERT model and UniLM framework to construct an interest privacy inference (IPI) model for interest recognition and to generate interpretable text that supports this interest. Finally, we constructed a dataset of 13,000 privacy-containing texts for experimentation. Experimental results show that the overall accuracy of the DPER model can reach 91.80%, while that of the IPI model can reach 98.3%. Simultaneously, we compare the proposed model with recent methods. The analysis of the results indicates that the proposed model exhibits better performance than previous methods.

Text localization and recognition of Chinese characters in natural scenes based on improved faster R-CNN

Text detection and recognition are widely used in daily life. Although it is a very rich market, it has very difficult challenges in practical application. The complex and changeable natural scenes lead to the complex background of the text in the image, which also reflects the research value of text information recognition and extraction in natural scenes. To solve the many problems faced by the recognition of Chinese characters, such as complex shapes and diverse structures, this paper uses VGG16 to extract features and introduces a two-layer bidirectional LSTM network. It improves Faster R-CNN by using a RPN to extract candidate boxes and adjust the position of candidate regions. In this paper, the improved model Faster BLSTM-CNN is tested, and the effectiveness experiment of feature extraction, the difference comparison before and after the improvement of the algorithm, and the comparison experiment with the traditional recognition algorithm are carried out respectively. And it finally carried out an experimental comparison of the combination of text recognition and positioning, and obtained the results. The algorithm Faster BLSTM-CNN in this paper is better in the localization and recognition of Chinese characters in the dataset. In the natural scene, the recognition rate of Faster BLSTM-CNN in this paper is 81.54%, the positioning accuracy is 88.14%, and the detection speed is 86 ms, which has improved performance. Therefore, the improvement of Faster R-CNN is effective. It can effectively locate and recognize Chinese characters in natural scenes.

A new deep CNN for 3D text localization in the wild through shadow removal

Text localization in the wild is challenging due to the presence of 2D and 3D texts, the presence of shadows, arbitrary orientated text with non-linear arrangements, varying lighting conditions as well as complex background. This paper proposes the first approach for 3D text localization in natural scene images through shadow removal and a new deep CNN model. In a first step, exploiting the observation that 3D text generates shadow information in natural scenes, the proposed model detects and removes the shadow pixels of 3D text based on the Generalized Gradient Vector Flow concept and a new clustering approach. The performance of the classification of 2D and 3D texts in the scene images is strengthened by using key features, including pixel strength, sharpness and edge potential, which are extracted to eliminate false text and shadow pixels. For text localization after removing shadow information, EfficientNet is used as an encoder (backbone) and UNet as a decoder in a novel way employing differential binarization. Experimental validation and comparative analysis with state-of-the-art approaches on both a new purpose-built dataset as well as on the benchmark datasets of ICDAR MLT 2019, ICDAR ArT 2019, CTW1500, DAST1500, Total-Text, and MSRATD500 for each of the different steps of the method, show that the proposed approach outperforms the existing methods.

Automated hand-marked semantic text recognition from photographs

Automated text recognition techniques have made significant advancements; however, certain tasks still present challenges. This study is motivated by the need to automatically recognize hand-marked text on construction defect tags among millions of photographs. To address this challenge, we investigated three methods for automating hand-marked semantic text recognition (HMSTR)—a modified scene text recognition-based (STR) approach, a two-step HMSTR approach, and a lumped approach. The STR approach involves locating marked text using an object detection model and recognizing it using a competition-winning STR model. Similarly, the two-step HMSTR approach first localizes the marked text and then recognizes the semantic text using an image classification model. By contrast, the lumped approach performs both localization and identification of marked semantic text in a single step using object detection. Among these approaches, the two-step HMSTR approach achieved the highest F1 score (0.92) for recognizing circled text, followed by the STR approach (0.87) and the lumped approach (0.78). To validate the generalizability of the two-step HMSTR approach, subsequent experiments were conducted using check-marked text, resulting in an F1 score of 0.88. Although the proposed methods have been tested specifically with tags, they can be extended to recognize marked text in reports or books.