Scholarlink

The proposed article is based on the intelligent system for the detection of defects in magnetic tiles using a transfer learning approach. The training models used in the study were already trained with various images so that the transfer learning approach could be efficiently utilized. The study uses a convolutional neural network (CNN) in multiple layers and processes data in a grid-like model. One main merit of using CNNs is that there is no need for preprocessing. To utilize the benefits of high speed, large bandwidth, and low latency 5G network is used. The standard dataset used in the article for experimental purposes, and the highest accuracy of 97.8% was obtained using the DeepTree model.

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 Learning-Based Detection of Surface Anomalies

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Rahul,
  • Ashwani Kharola,
  • Varun Pokhriyal,
  • Prajwal Dangwal

摘要

The proposed article is based on the intelligent system for the detection of defects in magnetic tiles using a transfer learning approach. The training models used in the study were already trained with various images so that the transfer learning approach could be efficiently utilized. The study uses a convolutional neural network (CNN) in multiple layers and processes data in a grid-like model. One main merit of using CNNs is that there is no need for preprocessing. To utilize the benefits of high speed, large bandwidth, and low latency 5G network is used. The standard dataset used in the article for experimental purposes, and the highest accuracy of 97.8% was obtained using the DeepTree model.