Scholarlink

<p>This research presents a novel real-time optical sensing-based image processing and deep learning method for 2D-3D human activity classification, specifically designed for virtual try-on systems. The proposed system utilizes advanced computer vision and machine learning techniques to accurately classify human activities in real-time, enabling immersive and interactive virtual try-on experiences. The virtual try-on system is a crucial component of e-commerce and retail technologies, allowing customers to try on virtual clothing and accessories in a realistic and interactive manner. However, existing virtual try-on systems often rely on manual input or simplistic algorithms, resulting in inaccurate or limited virtual try-on experiences.To address these limitations, this research proposes a real-time optical sensing-based approach, utilizing advanced image processing and deep learning techniques to accurately classify human activities and track body movements. The proposed system consists of three primary components: (1) real-time optical sensing, (2) image processing and feature extraction, and (3) deep learning-based human activity classification. The real-time optical sensing component utilizes advanced computer vision techniques to capture and process high-resolution images of the user's body movements. The image processing and feature extraction component extracts relevant features from the captured images, including pose, shape, and texture information. Finally, the deep learning-based human activity classification component utilizes a convolutional neural network (CNN) to classify the extracted features into predefined human activity categories. The proposed system is evaluated using a comprehensive dataset of human activities, including various body movements and poses. Experimental results demonstrate the effectiveness of the proposed system, achieving high accuracy and real-time performance. The proposed system has significant implications for virtual try-on systems, enabling immersive and interactive experiences for customers. Additionally, the proposed system can be applied to various other applications, including human–computer interaction, augmented reality, and healthcare.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A real time optical sensing-based image processing and deep learning method for 2D-3D human activity classification for virtual try-on systems

  • 期刊论文

  • 发表日期：

    2025年6月27日

查看全文
  • Akanksha Mrinali,
  • Pankaj Gupta

摘要

This research presents a novel real-time optical sensing-based image processing and deep learning method for 2D-3D human activity classification, specifically designed for virtual try-on systems. The proposed system utilizes advanced computer vision and machine learning techniques to accurately classify human activities in real-time, enabling immersive and interactive virtual try-on experiences. The virtual try-on system is a crucial component of e-commerce and retail technologies, allowing customers to try on virtual clothing and accessories in a realistic and interactive manner. However, existing virtual try-on systems often rely on manual input or simplistic algorithms, resulting in inaccurate or limited virtual try-on experiences.To address these limitations, this research proposes a real-time optical sensing-based approach, utilizing advanced image processing and deep learning techniques to accurately classify human activities and track body movements. The proposed system consists of three primary components: (1) real-time optical sensing, (2) image processing and feature extraction, and (3) deep learning-based human activity classification. The real-time optical sensing component utilizes advanced computer vision techniques to capture and process high-resolution images of the user's body movements. The image processing and feature extraction component extracts relevant features from the captured images, including pose, shape, and texture information. Finally, the deep learning-based human activity classification component utilizes a convolutional neural network (CNN) to classify the extracted features into predefined human activity categories. The proposed system is evaluated using a comprehensive dataset of human activities, including various body movements and poses. Experimental results demonstrate the effectiveness of the proposed system, achieving high accuracy and real-time performance. The proposed system has significant implications for virtual try-on systems, enabling immersive and interactive experiences for customers. Additionally, the proposed system can be applied to various other applications, including human–computer interaction, augmented reality, and healthcare.