SHShinobu Hasegawa

Shinobu Hasegawa

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Shinobu Hasegawa has a total of 46 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • EVALUATING NOVICE ANGKLUNG PLAYERS’ EMOTIONAL AND EXPERIENTIAL RESPONSES TO CUE-GUIDED SELF-TRAINING METHODS USING FACE RECOGNITION AND PARTICIPANT FEEDBACK

    by Bella Hardiyana, Hasegawa Shinobu, Diana Effendi, Beri Noviansyah, Zebiniso Khudoyshukurovna Abdunazarova

    This study investigates how novice angklung players emotionally and experientially respond to three cue-guided self-training methods: Notebar (NB), Hand-Sign Bot without Preview (HB), and Hand-Sign Bot with Preview (HBP). The purpose of the study is to clarify how different cue modalities influence comfort, focus, and engagement during traditional instrument learning. The novelty lies in combining continuous facial affect analysis using OpenFace with thematic analysis of open-ended participant feedback, allowing a detailed examination of cognitive and emotional demands during training. A total of twenty-seven novices, consisting of nine Indonesian and eighteen non-Indonesian participants, completed three counterbalanced training cycles. Emotional states were recorded through frame-by-frame valence and arousal extraction, and subjective perceptions were collected after each method. Results show that NB provides the most accessible and low-stress experience. HB is perceived as the most realistic but introduces the highest attentional load. HBP offers anticipatory support for some learners but creates split-attention difficulty for others. Indonesian participants generally exhibit more positive valence and lower arousal, indicating that cultural familiarity moderates affective responses. By integrating quantitative affective indicators with qualitative perceptions, this study identifies how cue design influences novice learning and offers guidance for developing adaptive and culturally responsive digital angklung training systems.

  • Generating Survey Hint-Net from Multiple Academic Papers with Cite-Clues

    by Jinghong Li, RuoZhu Sheng, Dawn Gu, Shinobu Hasegawa

    The exponential growth of academic literature has created significant challenges in tracking knowledge evolution and understanding research trajectories. While generative AI has shown promise in constructing knowledge graphs and questionanswering systems from academic papers, current approaches struggle to capture the nuanced citation relationships that reveal how knowledge flows and evolves. This research introduces 'cite-clues' - semantic patterns that reveal cite-motivations and knowledge transfer mechanisms between survey papers and regular papers. We propose Survey Hint-net, a novel visualization framework that leverages these cite-clues to map knowledge evolution pathways in reasoning explanation. This framework enables researchers to navigate from foundational concepts to specific solutions and track the development of research topics over time.

  • Open Access

    Non-Semantic Multimodal Fusion for Predicting Segment Access Frequency in Lecture Archives

    by Ruozhu Sheng, Jinghong Li, Shinobu Hasegawa

    This study proposes a non-semantic multimodal approach to predict segment access frequency (SAF) in lecture archives. Such archives, widely used as supplementary resources in modern education, often consist of long, unedited recordings that are difficult to navigate and review efficiently. The predicted SAF, an indicator of student viewing behavior, serves as a practical proxy for student engagement. The increasing volume of recorded material renders manual editing and annotation impractical, making the automatic identification of high-SAF segments crucial for improving accessibility and supporting targeted content review. The approach focuses on lecture archives from a real-world blended learning context, characterized by resource constraints such as no specialized hardware and limited student numbers. The model integrates multimodal features from instructor’s actions (via OpenPose and optical flow), audio spectrograms, and slide page progression—a selection of features that makes the approach applicable regardless of lecture language. The model was evaluated on 665 labeled one-minute segments from one such course. Experiments show that the best-performing model achieves a Pearson correlation of 0.5143 in 7-fold cross-validation and 61.05% average accuracy in a downstream three-class classification task. These results demonstrate the system’s capacity to enhance lecture archives by automatically identifying key segments, which aids students in efficient, targeted review and provides instructors with valuable data for pedagogical feedback.

  • Open Access

    Multimodal Non-Semantic Feature Fusion for Predicting Segment Access Frequency in Lecture Archives

    by Ruozhu Sheng, Jinghong Li, Shinobu Hasegawa

    This study proposes a multimodal neural network-based approach to predict segment access frequency in lecture archives. These archives, widely used as supplementary resources in modern education, often consist of long, unedited recordings that make it difficult to keep students engaged. Captured directly from face-to-face lectures without post-processing, they lack visual appeal. Meanwhile, the increasing volume of recorded material renders manual editing and annotation impractical. Automatically detecting high-engagement segments is thus crucial for improving accessibility and maintaining learning effectiveness. Our research focuses on real classroom lecture archives, characterized by unedited footage, no additional hardware (e.g., eye-tracking), and limited student numbers. We approximate student engagement using segment access frequency as a proxy. Our model integrates multimodal features from teachers' actions (via OpenPose and optical flow), audio spectrograms, and slide page progression. These features are deliberately chosen for their non-semantic nature, making the approach applicable regardless of lecture language. Experiments show that our best model achieves a Pearson correlation of 0.5143 in 7-fold cross-validation and 69.32 percent average accuracy in a downstream three-class classification task. The results, obtained with high computational efficiency and a small dataset, demonstrate the practical feasibility of our system in real-world educational contexts.

  • Open Access

    Development and Evaluation of a Low-Jitter Hand Tracking System for Improving Typing Efficiency in a Virtual Reality Workspace

    by Tianshu Xu, Wen Gu, Koichi Ota, Shinobu Hasegawa

    Virtual reality technology promises to transform immersive experiences across various applications, particularly within office environments. Despite its potential, the challenge of achieving efficient text entry in virtual reality persists. This study addresses this obstacle by introducing a novel machine learning-based solution, namely, the two-stream long short-term memory typing method, to enhance text entry performance in virtual reality. The two-stream long short-term memory method utilizes the back-of-the-hand image, employing a long short-term memory network and a Kalman filter to enhance hand position tracking accuracy and minimize jitter. Through statistical analysis of the data collected in the experiment and questionnaire results, we confirmed the effectiveness of the proposed method. In addition, we conducted an extra experiment to explore the differences in users’ typing behavior between regular typing and virtual reality-based typing. This additional experiment provides valuable insights into how users adapt their typing behavior in different environments. These findings represent a significant step in advancing text entry within virtual reality, setting the stage for immersive work experiences in office environments and beyond.


  • Open Access

    The Impact of a New Specialized Learning Platform on Enhancing Students’ Skills in Designing and Producing Online Educational Talking Books

    by Mohamed Elsayed Ahmed, Shinobu Hasegawa

    The use of digital audio-based learning applications in classrooms is increasing. Thus, instructional technology professionals need to become more skilled in designing and developing these apps. To succeed in the future learning process, students studying instructional technology need to have mastered these skills. This study aims to examine the impact of a specialized platform proposed to cultivate students’ skills in easily designing and producing online educational talking books (online ETBs). To improve the target abilities, we developed a new specialized platform as a web-based system for this investigation that includes several innovative tools and special learning content. The research design used was quasi-experimental. Thirty-two undergraduate students majoring in instructional technology who were in the third grade examined the proposed platform. With the pre-and post-test design, all participants were placed into one group. Three online instruments were used to gather the data: a performance observation card; a product evaluation card; and a usability questionnaire. According to the results, the suggested platform considerably increased the practical skills required to design and create an online ETB. Additionally, the proposed platform positively impacted the production of online ETB products with an acceptable proficiency level across various educational subjects. The customized platform has also proven to be a comfortable learning environment for the participating instructional technology students. This platform may be utilized as an ICT teaching tool to enhance instructional technology departments as it is appropriate for constructing online ETBs in various educational areas without coding barriers.


  • Open Access

    A Proposal for a Quantitative Evaluation Model for Error Image Generation in L2 Vocabulary Learning

    by Kazuki Sugita, Wen Gu, Koichi Ota, Prarinya Siritanawan, Shinobu Hasegawa

    Vocabulary learning that incorporates visual information has become widely recognized as an alternative to context-based methods. However, few studies focus on learners' incorrect answers. On the Other hand, fossilization caused by repeated errors has been a concern. Our proposed system, L-VEIGe, effectively prevents repeated errors by visualizing learners' incorrect answers through image generation, which encourages introspection. However, there exists a 'Feature Disappearance' problem, where the generated images for incorrect answers lack sufficient information for comprehension. This study proposes a method for quantitatively evaluating these error images from a cognitive perspective.