The detection and acknowledgment of logos holds significant importance in the corporate sphere, facilitating the detection of unauthorized logo usage and ensuring trademark uniqueness within specific industry sectors. Presently, convolutional neural networks powered by deep learning are widely utilized for image recognition. However, their effectiveness is dependent on a substantial volume of training images which may not always be readily available. This study suggests employing Test Time Augmentation to address dataset constraints by expanding the original dataset, thereby enhancing classification accuracy and preventing overfitting. Test-Time Augmentation is a method used to improve the accuracy of convolutional neural networks by creating numerous augmented variations of the test images and then merging their predictions. The research findings indicate that the application of TTA has the highest performance on the VGG16 model with 98% precision, 99% recall, and 98% F1-score, and 98.87% accuracy
24 June 2025 in Preservation, Digital Technology & Culture
Batik, a traditional textile art form, holds profound cultural and historical value in Indonesia, where its motifs are rich in symbolic meaning. As efforts to digitally preserve batik grow, there is an increasing need for accurate and scalable classification methods to analyze its intricate patterns. Manual classification remains labor-intensive and prone to subjectivity. This study investigates the use of machine learning and computer vision techniques to automate batik motif classification. By leveraging deep feature extraction through the VGG16 architecture and integrating it with machine learning classifiers such as Support Vector Machines (SVM), Multi-Layer Perceptron (MLP), Convolutional Neural Networks (CNN), and XGBoost, the study aims to improve classification performance. Experimental results demonstrate that the VGG16+XGBoost model achieves the highest performance with an average accuracy of 89.83 % across five-fold cross-validation, outperforming standalone models including SVM (64.29 %), CNN (65.76 %), and XGBoost without deep features (64.87 %). Meanwhile, the retrained VGG16 model attained an accuracy of 87.77 %, confirming the benefit of combining deep and traditional learning approaches. These findings highlight the effectiveness of deep feature extraction in capturing the fine-grained textures of batik motifs and demonstrate the potential of such hybrid approaches in supporting the digital preservation and classification of traditional cultural artifacts.
30 August 2024 in Journal of Social Sciences and Technology for Community Service (JSSTCS)
Artificial intelligence (AI) has emerged as one of the most transformative and influential technological advancements, significantly impacting sectors like healthcare, finance, and education in the rapidly evolving digital era. However, vocational high school students often have limited knowledge about AI, primarily due to the superficial coverage of AI in their curriculum and the lack of resources and trained educators. This study aimed to assess the effectiveness of an intervention designed to enhance AI awareness and understanding among vocational high school students. A comprehensive program was implemented, involving initial assessments, interactive face-to-face training sessions, and facilitated group discussions. Descriptive analysis using statistical methods was employed to compare pretest and posttest data, revealing significant improvements in students' knowledge and attitudes towards AI. For example, the average score for the question "I know what Artificial Intelligence is" increased from 1.77 to 2.93, while "I can name an example of an AI application" rose from 1.76 to 2.85. Additionally, students showed increased awareness of AI's impact on future careers and a greater interest in learning more about AI, with scores rising from 2.20 to 2.85 and 2.48 to 2.96, respectively. The intervention proved effective in significantly enhancing students' understanding and attitudes towards AI, highlighting the value of targeted educational programs. By addressing gaps in AI education and fostering a proactive learning environment, the approach equips students with the knowledge and skills necessary to thrive in an AI-driven job market. This comprehensive strategy not only prepares students for future career challenges but also empowers them to leverage AI technology for career advancement.
25 August 2026 in Preservation, Digital Technology & Culture
Traditional board games are increasingly digitized, yet many digital adaptations reduce embodied interaction and weaken cultural meaning. This study proposes a multimodal framework to revitalize Dakon as an intangible cultural heritage practice by integrating gesture-based computer vision and large language models. The system combines MediaPipe hand landmark detection, a Support Vector Machine classifier for gesture recognition, a deterministic Dakon game engine, and event-driven generative dialogue using Gemini 3 Flash. Experimental results show perfect precision across active gesture classes, with recall values ranging from 0.72 to 0.77, indicating that performance limitations primarily stem from hand detection robustness rather than classification accuracy. User experience evaluation involving twenty participants demonstrates high levels of engagement, perceived authenticity, clarity, and responsiveness. In addition, linguistically informed evaluation of 120 LLM-generated responses by expert evaluators reveals strong performance in cultural relevance, contextual accuracy, and motivational quality. The findings highlight that separating deterministic game logic from generative AI preserves gameplay authenticity while enabling contextual cultural interpretation. This study contributes a heritage-centered design principle in which generative AI functions as an interpretative layer rather than a rule-modifying agent. The proposed approach extends digital cultural preservation beyond static digitization toward interactive, embodied, and participatory experiences.