Scholarlink

This study delves into the intersection of art and artificial intelligence, focusing particularly on artist recognition using deep learning techniques. Initially, we curate a dataset comprising artworks from eleven renowned artists. To address data imbalances, we employ class weights and meticulous data augmentation. Our modeling approach centers on the ResNet50 architecture, with a special emphasis on training shallow layers to capture the essence of creativity effectively. Achieving nearly 99% accuracy on the training set and 85% on the cross-validation set, the final model marks a significant advancement in robots’ capacity to identify and appreciate artistic masterpieces. This research not only redefines the relationship between human and machine creativity but also paves the way for exploring avenues at the convergence of art and technology. Our result shows that the ResNet50 is a powerful method for artist identification and able to learn a representation of painting.

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

Artist Prediction of Painting Through Transfer Learning

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Bharat Singh,
  • Pranay Reddy,
  • Nidhi Kushwaha

摘要

This study delves into the intersection of art and artificial intelligence, focusing particularly on artist recognition using deep learning techniques. Initially, we curate a dataset comprising artworks from eleven renowned artists. To address data imbalances, we employ class weights and meticulous data augmentation. Our modeling approach centers on the ResNet50 architecture, with a special emphasis on training shallow layers to capture the essence of creativity effectively. Achieving nearly 99% accuracy on the training set and 85% on the cross-validation set, the final model marks a significant advancement in robots’ capacity to identify and appreciate artistic masterpieces. This research not only redefines the relationship between human and machine creativity but also paves the way for exploring avenues at the convergence of art and technology. Our result shows that the ResNet50 is a powerful method for artist identification and able to learn a representation of painting.