Scholarlink

With the rapid growth of e-commerce and the increasing importance of visual content, the development of advanced fashion recommendation systems has become paramount. This work presents a fresh methodology to fashion recommendations leveraging cutting-edge deep learning models for feature extraction and the K-Nearest Neighbours (KNN) algorithm for similarity-based image retrieval. The proposed system employs pre-trained convolutional neural networks (CNNs) such as ResNet50 and Mobile Net to extract rich and discriminative features from fashion images. All these features are summing up high-situation representations of different fashion attributes to allow for an adequate understanding of the visual characteristics presented in the input images. Afterward, a K-Nearest Neighbours algorithm is used to find out similar fashion details from a database of images measures how similar the query image is to other images in the database based on their rooted features. Furthermore, it also sorts and displays top 5 fashion pictures that share almost the same visual attributes as the query one. The integration of deep literacy for point birth and KNN algorithm guarantees that this recommendation system has the ability to make personalized and visually appealing suggestions on what to wear. This recommended system improves a stoner’s experience of learning about and trying various fashion options by not only considering particular fashion pieces but rather whole looks or styles associated with them. Experimental findings demonstrate the efficiency of this proven system which performs accurate and significant recommendations about materials used in making clothes.

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

Deep Feature Extraction for Fashionable Fabrics: Using ResNet50, MobileNet, and CNN

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • V. V. Krishna Reddy,
  • Susmitha Cherukuri,
  • Karthikeya Vanaparla,
  • Lakshman Rishi Avula

摘要

With the rapid growth of e-commerce and the increasing importance of visual content, the development of advanced fashion recommendation systems has become paramount. This work presents a fresh methodology to fashion recommendations leveraging cutting-edge deep learning models for feature extraction and the K-Nearest Neighbours (KNN) algorithm for similarity-based image retrieval. The proposed system employs pre-trained convolutional neural networks (CNNs) such as ResNet50 and Mobile Net to extract rich and discriminative features from fashion images. All these features are summing up high-situation representations of different fashion attributes to allow for an adequate understanding of the visual characteristics presented in the input images. Afterward, a K-Nearest Neighbours algorithm is used to find out similar fashion details from a database of images measures how similar the query image is to other images in the database based on their rooted features. Furthermore, it also sorts and displays top 5 fashion pictures that share almost the same visual attributes as the query one. The integration of deep literacy for point birth and KNN algorithm guarantees that this recommendation system has the ability to make personalized and visually appealing suggestions on what to wear. This recommended system improves a stoner’s experience of learning about and trying various fashion options by not only considering particular fashion pieces but rather whole looks or styles associated with them. Experimental findings demonstrate the efficiency of this proven system which performs accurate and significant recommendations about materials used in making clothes.