Search

Vocabulary acquisition and retention is an essential part of learning a foreign language and many learners use flashcard applications to repetitively increase vocabulary retention. However, it can be difficult for learners to remember new words and phrases without any context. In this paper, we propose a system that allows users to acquire new vocabulary with media which gives context to the words. Theoretically, this use of multimedia context should enable users to practice with interest and increased motivation, which has been shown to enhance the effects of contextual language learning. An experiment with 46 English as foreign language learners showed better retention after two weeks with the proposed system as compared to ordinary flashcards. However, the impact was not universally beneficial to all learners. An analysis of participant attributes that were gathered through surveys and questionnaires shows a link between personality and learning traits and affinity for learning with this system. This result indicates that the proposed system provides a significant advantage in vocabulary retention for some users, while other users should stay with traditional flashcard applications. The implications of this study indicate the need for the development of more personalized learning applications. [ABSTRACT FROM AUTHOR]

Attention-based encoder–decoder scene text recognition (STR) architectures have been proven effective in recognizing text in the real world, thanks to their ability to learn an internal language model. Nevertheless, the cross-attention operation that is used to align visual and linguistic features during decoding is computationally expensive, especially in low-resource environments. To address this bottleneck, we propose a cross-attention-free STR framework that still learns a language model. The framework we propose is ViTSTR-Transducer, which draws inspiration from ViTSTR, a vision transformer (ViT)-based method designed for STR and the recurrent neural network transducer (RNN-T) initially introduced for speech recognition. The experimental results show that our ViTSTR-Transducer models outperform the baseline attention-based models in terms of the required decoding floating point operations (FLOPs) and latency while achieving a comparable level of recognition accuracy. Compared with the baseline context-free ViTSTR models, our proposed models achieve superior recognition accuracy. Furthermore, compared with the recent state-of-the-art (SOTA) methods, our proposed models deliver competitive results. [ABSTRACT FROM AUTHOR]

Connectionist temporal classification (CTC) is a favored decoder in scene text recognition (STR) for its simplicity and efficiency. However, most CTC-based methods utilize one-dimensional (1D) vector sequences, usually derived from a recurrent neural network (RNN) encoder. This results in the absence of explainable 2D spatial relationship between the predicted characters and corresponding image regions, essential for model explainability. On the other hand, 2D attention-based methods enhance recognition accuracy and offer character location information via cross-attention mechanisms, linking predictions to image regions. However, these methods are more computationally intensive, compared with the 1D CTC-based methods. To achieve both low latency and model explainability via character localization using a 1D CTC decoder, we propose a marginalization-based method that processes 2D feature maps and predicts a sequence of 2D joint probability distributions over the height and class dimensions. Based on the proposed method, we newly introduce an association map that aids in character localization and model prediction explanation. This map parallels the role of a cross-attention map, as seen in computationally-intensive attention-based architectures. With the proposed method, we consider a ViT-CTC STR architecture that uses a 1D CTC decoder and a pretrained vision Transformer (ViT) as a 2D feature extractor. Our ViT-CTC models were trained on synthetic data and fine-tuned on real labeled sets. These models outperform the recent state-of-the-art (SOTA) CTC-based methods on benchmarks in terms of recognition accuracy. Compared with the baseline Transformer-decoder-based models, our ViT-CTC models offer a speed boost up to 12 times regardless of the backbone, with a maximum 3.1% reduction in total word recognition accuracy. In addition, both qualitative and quantitative assessments of character locations estimated from the association map align closely with those from the cross-attention map and ground-truth character-level bounding boxes. [ABSTRACT FROM AUTHOR]

Online knowledge-based peer-production communities, like question and answer sites (Q&A), often rely on gamification, e.g., through reputation points, to incentivize users to contribute frequently and effectively. These gamification techniques are important for achieving the critical mass that sustains a community and enticing new users to join. However, aging communities tend to build "poverty traps" that act as barriers for new users. In this paper, we present our investigation of 32 domain communities from Stack Exchange and our analysis of how different subjects impact the development of early user advantage. Our results raise important questions about the accessibility of knowledge-based peer-production communities. We consider the analysis results in the context of changing information needs and the relevance of Q&A in the future. Our findings inform policy design for building more equitable knowledge-based peer-production communities and increasing the accessibility to existing ones. [ABSTRACT FROM AUTHOR]

Wearable devices offer a wealth of data for ubiquitous computing researchers. For instance, sleep data from a wearable could be used to identify an individual's harmful habits. Recently, devices which are unobtrusive in size, setup, and maintenance are becoming commercially available. However, most data validation for these devices come from brief, short-term laboratory studies or experiments which have unrepresentative samples that are also inaccessible to most researchers. For wearables research conducted in-the-wild, the prospect of running a study has the risk of financial costs and failure. Thus, when researchers conduct in-the-wild studies, the majority of participants tend to be university students. In this paper, we present a month-long in-the-wild study with 31 Japanese adults who wore a sleep tracking device called the Oura ring. The high device usage results found in this study can be used to inform the design and deployment of longer-term mid-size in-the-wild studies. [ABSTRACT FROM AUTHOR]

Arbuscular mycorrhizal fungi (AMF) infect plant roots and are hypothesized to improve plant growth. Recently, AMF is now available for axenic culture. Therefore, AMF is expected to be used as a microbial fertilizer. To evaluate the usefulness of AMF as a microbial fertilizer, we need to investigate the relationship between the degree of root colonization of AMF and plant growth. The method popularly used for calculation of the degree of root colonization, termed the magnified intersections method, is performed manually and is too labor-intensive to enable an extensive survey to be undertaken. Therefore, we automated the magnified intersections method by developing an application named "Tool for Analyzing root images to calculate the Infection rate of arbuscular Mycorrhizal fungi: TAIM." TAIM is a web-based application that calculates the degree of AMF colonization from images using automated computer vision and pattern recognition techniques. Experimental results showed that TAIM correctly detected sampling areas for calculation of the degree of infection and classified the sampling areas with 87.4% accuracy. TAIM is publicly accessible at http://taim.imlab.jp/. [ABSTRACT FROM AUTHOR]

The observer effect found in laboratory studies has long posed a problem for researchers. In-the-wild studies reduce the observer effect, but have problems with gathering accurately labeled data usable for training algorithms. Manual labeling is time-consuming, obtrusive, and unfeasible, and if done by the researchers, it potentially violates the privacy of the participants. In this article, we present a labeling workflow based on an in-the-wild study that investigated cognitive state changes through eye-gaze in naturalistic settings. We contribute a setup that enables participants to label their data unobtrusively and quickly. We use J!NS MEME electrooculography glasses, Narrative Clip 2 wearable cameras, and a proprietary data tagging software package. Our setup is reproducible for field studies, preserves data integrity, and maintains participant privacy. This workflow can be extended to other studies in pervasive and ubiquitous computing and is especially suitable for deployment in the pandemic and postpandemic world. [ABSTRACT FROM AUTHOR]

The automated generation of fonts containing a large number of characters is in high demand. For example, a typical Japanese font requires over 1000 characters. Unfortunately, professional typographers create the majority of fonts, resulting in significant financial and time investments for font generation. The main contribution of this article is the development of a method that automatically generates a target typographic font containing thousands of characters, from a small subset of character images in the target font. We generate characters other than the subset so that a complete font is obtained. We propose a novel font generation method with the capability to deal with various fonts, including a font composed of distinctive strokes, which are difficult for existing methods to handle. We demonstrated the proposed method by generating 2965 characters in 47 fonts. Moreover, objective and subjective evaluations verified that the generated characters are similar to the original characters. [ABSTRACT FROM AUTHOR]

The integration of ubiquitous technologies in the field of education has considerably enhanced our way of learning. Such technologies enable students to get a gradual feedback about their performance and to provide adapted learning materials. It is particularly important in the domain of foreign language learning which requires intense daily practice. One of the main inputs of adaptive learning systems is the user’s understanding of a reading material. The reader’s understanding can be divided into two parts: the objective understanding and the subjective understanding. The objective understanding can be measured by comprehension questions about the content of the text. The subjective understanding is the reader’s perception of his own understanding. The subjective understanding plays an important role in the reader’s motivation, self-esteem and confidence. However, its automatic estimation remains a challenging task. This paper is one of the first to propose a method to estimate the subjective understanding. We show that using the eye gaze to predict the subjective understanding improves the estimation by 13% as compared to using comprehension questions. [ABSTRACT FROM AUTHOR]

Applying mobile sensing technology to cognitive tasks will enable novel forms of activity recognition. [ABSTRACT FROM PUBLISHER]

Recognizing characters in scene images suffering from perspective distortion is a challenge. Although there are some methods to overcome this difficulty, they are time-consuming. In this paper, we propose a set of affine-invariant features and a new recognition scheme called 'progressive removing' that can help reduce the processing time. Progressive removing gradually removes less feasible categories and skew angles by using multiple classifiers. We observed that progressive removing and the use of the affine invariant features reduced the processing time by about 60% in comparison to a trivial algorithm without decreasing the recognition rate. © 2012 Wiley Periodicals, Inc. Electr Eng Jpn, 180(2): 55-63, 2012; Published online in Wiley Online Library (). DOI 10.1002/eej.22276 [ABSTRACT FROM AUTHOR]

In this paper, we propose a system for recommending tags for images. The proposed method presents to the user various tags with high precision by taking into account both the co-occurrence of tags and tags for similar images. Additional searches with user feedback enable us to find some new tags relevant to the image of interest. In order to test the proposed method, we compared it with a method based only on the co-occurrence of tags, as well as a method based only on the similarity of images. From the experimental results using 16,002 images, we have confirmed that the proposed method is capable of finding more tags compared to the methods used for comparison. We have also confirmed that the additional search is effective for finding additional tags. © 2011 Wiley Periodicals, Inc. Electron Comm Jpn, 94(12): 57-64, 2011; Published online in Wiley Online Library (). DOI 10.1002/ecj.10342 [ABSTRACT FROM AUTHOR]

In order to make collaborative business activities fruitful, it is essential to know characteristics of organizations and persons in more details and to gather information relevant to the activities. In this paper, we describe a notion of “information recycling” that actualizes these requirements by analyzing documents. The key of recycling information is to utilize annotations on documents as clues for generating user profiles and for weighting contents in the context of the activities. We also propose a method of extracting annotations on paper documents just by pressing one button with the help of techniques of camera-based document image analysis. Experimental results demonstrate that it is fundamentally capable of acquiring annotations on paper documents on the condition that their electronic versions without annotations are available for the processing. © 2008 Wiley Periodicals, Inc. Electr Eng Jpn, 165(2): 60–68, 2008; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/eej.20516 [ABSTRACT FROM AUTHOR]

Page segmentation is a process used to extract such components as columns, figures, tables, and photos from an image of a document. This article proposes a page segmentation technique that is stable, irrespective of component shape or tilted document image, based on analyzing the white region (background) of the document image. When we process a document that has non-rectangular and tilted components, the boundary of the components, that is, the white region, takes any shape. Thus, important questions include how to express white regions and how to process them. The proposed method uses thin lines that are extracted by thinning as an expression of white regions. Based on this expression of white regions, page segmentation is defined as extracting loops that surround the components. The proposed method extracts loops by eliminating unnecessary thin lines, for example, those that represent line spacing and character spacing. We try to use not only the feature of white regions, but also those of black regions, and to process several kinds of document layout. This paper examines the effectiveness and limitations of the proposed method based on experimental results that are taken from 80 sample images that are tilted from 0 to 45 degrees. © 1998 Scripta Technica. Syst Comp Jpn, 29(3): 59–68, 1998 [ABSTRACT FROM AUTHOR]

Post-processing of character recognition refers to the processing used to correct the errors in character recognition. When the input is a string representing a sentence in the highly precise error correction, it is desired that the syntactic as well as semantic examinations should be made at the sentence level. This paper assumes that the morphemes, syntax and semantics of the input sentence can be analyzed, and proposes a method that uses the syntactic and semantic analysis in the post-processing. The proposed method receives the list of candidate characters up to the fifth, and outputs the sentence that is adequate from the viewpoints of both syntax and semantics. The method features the following three points: (1) in word matching, it is examined also whether or not a sentence adequate from the viewpoints of syntax and semantics can be composed, and then the inadequate words extraction is inhibited; (2) characters having stronger syntactic and semantic constraints, such as the single-character particle and the conjugational suffix, are estimated top-down. Then, the case where the adequate character is not contained in the candidates can be handled; and (3) the words for which the adequateness cannot be determined from the syntactic or semantic viewpoint are selected by character re-recognition processing. An experiment is executed (or 50 sample sentences. The character recognition rate is improved from 83.0 percent to 98.0 percent, and the sentence recognition rate is improved from 10.0 percent to 94.0 percent. Compared to the method based only on word matching, the sentence recognition rate is improved by more than 20 percent. In other words, the effectiveness of the proposed method is demonstrated. [ABSTRACT FROM AUTHOR]

Image binarization is a task to convert gray-level images into bi-level ones. Its underlying notion can be simply thought of as threshold selection. However, the result of binarization will cause significant influence on the process of image recognition or understanding. In this paper we discuss a new binarization method, named CMB (connectionist model binarization), which uses the connectionist model. In the method a gray-level histogram is input to a multilayer network trained with the back-propagation algorithm to obtain a threshold which gives a visually suitable binarized image. From the experimental results, it was verified that CMB is an effective binarization method in comparison with other methods. [ABSTRACT FROM AUTHOR]

SUMMARY Specific object recognition methods based on the exact matching of feature vectors are known as methods that can achieve high recognition performance for large-scale three-dimensional specific object recognition. Since there are few common three-dimensional object datasets whose size is sufficient to explore the effect of differences in object dataset composition and the effect of increasing number of objects for recognition, these effects for specific object recognition methods based on exact matching of feature vectors have been discussed insufficiently. The number of objects in well-known datasets (e.g., COIL-100) is around 100. Therefore, in this research, we prepared a dataset of 1002 three-dimensional objects by themselves. In this paper, we discuss the effect of dataset differences, which are based on object structure, texture, and the number of objects, for methods such as that based on the Bloomier filter and that based on a hash table with this dataset in addition to COIL-100. © 2013 Wiley Periodicals, Inc. Electron Comm Jpn, 96(9): 33-45, 2013; Published online in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/ecj.11414 [ABSTRACT FROM AUTHOR]

To study plant organs, it is necessary to investigate the three-dimensional (3D) structures of plants. In recent years, non-destructive measurements through computed tomography (CT) have been used to understand the 3D structures of plants. In this study, we use the Chrysanthemum seticuspe capitulum inflorescence as an example and focus on contact points between the receptacles and florets within the 3D capitulum inflorescence bud structure to investigate the 3D arrangement of the florets on the receptacle. To determine the 3D order of the contact points, we constructed slice images from the CT volume data and detected the receptacles and florets in the image. However, because each CT sample comprises hundreds of slice images to be processed and each C. seticuspe capitulum inflorescence comprises several florets, manually detecting the receptacles and florets is labor-intensive. Therefore, we propose an automatic contact point detection method based on CT slice images using image recognition techniques. The proposed method improves the accuracy of contact point detection using prior knowledge that contact points exist only around the receptacle. In addition, the integration of the detection results enables the estimation of the 3D position of the contact points. According to the experimental results, we confirmed that the proposed method can detect contacts on slice images with high accuracy and estimate their 3D positions through clustering. Additionally, the sample-independent experiments showed that the proposed method achieved the same detection accuracy as sample-dependent experiments., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2024 Matsumoto, Utsumi, Kozuka, Iwamura, Nakai, Yamauchi, Karahara, Mineyuki, Hoshino, Uesugi and Kise.)

Manual segmentation of the petals of flower computed tomography (CT) images is time-consuming and labor-intensive because the flower has many petals. In this study, we aim to obtain a three-dimensional (3D) structure of Camellia japonica flowers and propose a petal segmentation method using computer vision techniques. Petal segmentation on the slice images fails by simply applying the segmentation methods because the shape of the petals in CT images differs from that of the objects targeted by the latest instance segmentation methods. To overcome these challenges, we crop two-dimensional (2D) long rectangles from each slice image and apply the segmentation method to segment the petals on the images. Thanks to cropping, it is easier to segment the shape of the petals in the cropped images using the segmentation methods. We can also use the latest segmentation method for the task because the number of images used for training is augmented by cropping. Subsequently, the results are integrated into 3D to obtain 3D segmentation volume data. The experimental results show that the proposed method can segment petals on slice images with higher accuracy than the method without cropping. The 3D segmentation results were also obtained and visualized successfully., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2024 Naka, Utsumi, Iwamura, Tsukaya and Kise.)

Flowers exhibit morphological diversity in the number and positional arrangement of their floral organs, such as petals. The petal arrangements of blooming flowers are represented by the overlap position relation between neighboring petals, an indicator of the floral developmental process; however, only specialists are capable of the petal arrangement identification. Therefore, we propose a method to support the estimation of the arrangement of the perianth organs, including petals and tepals, using image recognition techniques. The problem for realizing the method is that it is not possible to prepare a large number of image datasets: we cannot apply the latest machine learning based image processing methods, which require a large number of images. Therefore, we describe the tepal arrangement as a sequence of interior-exterior patterns of tepal overlap in the image, and estimate the tepal arrangement by matching the pattern with the known patterns. We also use methods that require less or no training data to implement the method: the fine-tuned YOLO v5 model for flower detection, GrubCut for flower segmentation, the Harris corner detector for tepal overlap detection, MAML-based interior-exterior estimation, and circular permutation matching for tepal arrangement estimation. Experimental results showed good accuracy when flower detection, segmentation, overlap location estimation, interior-exterior estimation, and circle permutation matching-based tepal arrangement estimation were evaluated independently. However, the accuracy decreased when they were integrated. Therefore, we developed a user interface for manual correction of the position of overlap estimation and interior-exterior pattern estimation, which ensures the quality of tepal arrangement estimation., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2024 Nakatani, Utsumi, Fujimoto, Iwamura and Kise.)

This paper describes a method based on a deep neural network (DNN) for estimating the number of tillers on a plant. A tiller is a branch on a grass plant, and the number of tillers is one of the most important determinants of yield. Traditionally, the tiller number is usually counted by hand, and so an automated approach is necessary for high-throughput phenotyping. Conventional methods use heuristic features to estimate the tiller number. Based on the successful application of DNNs in the field of computer vision, the use of DNN-based features instead of heuristic features is expected to improve the estimation accuracy. However, as DNNs generally require large volumes of data for training, it is difficult to apply them to estimation problems for which large training datasets are unavailable. In this paper, we use two strategies to overcome the problem of insufficient training data: the use of a pretrained DNN model and the use of pretext tasks for learning the feature representation. We extract features using the resulting DNNs and estimate the tiller numbers through a regression technique. We conducted experiments using side-view whole plant images taken with plan backgroud. The experimental results show that the proposed methods using a pretrained model and specific pretext tasks achieve better performance than the conventional method., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2023 Kinose, Utsumi, Iwamura and Kise.)