Scholarlink

Over the past years, automatic image colouring processes have generated considerable interest in several applications, including the restoration of old or damaged images we aim to achieve a comprehensive and adaptable colourization process by leveraging a combination of architectural components. Specifically, we employ a CGAN augmented with ResNets and U-Net architectures, incorporating the leaky ReLU activation function within a GAN framework. This approach offers improved performance, particularly in scenarios involving highly processed or thematically specialized photographs.

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

Colourization of Greyscale Images Using GAN with ResNet as the Backbone

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Shikhar Gaur,
  • Md Salman,
  • Anshu Khurana,
  • Seema Kalonia

摘要

Over the past years, automatic image colouring processes have generated considerable interest in several applications, including the restoration of old or damaged images we aim to achieve a comprehensive and adaptable colourization process by leveraging a combination of architectural components. Specifically, we employ a CGAN augmented with ResNets and U-Net architectures, incorporating the leaky ReLU activation function within a GAN framework. This approach offers improved performance, particularly in scenarios involving highly processed or thematically specialized photographs.