Scholarlink

<p>With the growing influence of digital commerce and artificial intelligence (AI), the fashion industry is undergoing a paradigm shift, predominantly in the domain of personalized recommendation systems. Traditional recommendation systems are dependent on historical purchase data and often fail to take rapidly evolving fashion trends into consideration. This paper presents a trend-inspired style matching model that uses deep learning (DL) and AI techniques to recommend outfits aligned with current fashion trends and user style preferences. A dataset comprising of 250 fashion images was curated from online sources, and four state-of-the-art deep learning models, VGG16, ResNet50, InceptionV3, and EfficientNetB0, were employed for feature extraction. Sketches and fashion images were preprocessed, and similarity between feature vectors was computed using cosine similarity, followed by clustering based ranking to identify the closest matching styles. The results demonstrate that EfficientNetB0 and InceptionV3 offer superior performance in terms of feature separability and recommendation accuracy compared to other models. The integration of AI, trend analysis, and visual similarity computation offers a robust framework for generating relevant and personalized fashion recommendations, addressing the limitations of traditional methods and enhancing user experience in fashion retail platforms.</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

AI-Based Fashion Recommendations Aligned with Current Trends from Sketch to Style

  • 期刊论文

  • 发表日期：

    2025年12月19日

查看全文
  • A. S. Niveditha,
  • R. Subha,
  • M. Selvadass

摘要

With the growing influence of digital commerce and artificial intelligence (AI), the fashion industry is undergoing a paradigm shift, predominantly in the domain of personalized recommendation systems. Traditional recommendation systems are dependent on historical purchase data and often fail to take rapidly evolving fashion trends into consideration. This paper presents a trend-inspired style matching model that uses deep learning (DL) and AI techniques to recommend outfits aligned with current fashion trends and user style preferences. A dataset comprising of 250 fashion images was curated from online sources, and four state-of-the-art deep learning models, VGG16, ResNet50, InceptionV3, and EfficientNetB0, were employed for feature extraction. Sketches and fashion images were preprocessed, and similarity between feature vectors was computed using cosine similarity, followed by clustering based ranking to identify the closest matching styles. The results demonstrate that EfficientNetB0 and InceptionV3 offer superior performance in terms of feature separability and recommendation accuracy compared to other models. The integration of AI, trend analysis, and visual similarity computation offers a robust framework for generating relevant and personalized fashion recommendations, addressing the limitations of traditional methods and enhancing user experience in fashion retail platforms.