Scholarlink

Image classification is a crucial task in computer vision, with significant applications in fields such as healthcare diagnostics, autonomous navigation, and security systems. This paper introduces a hybrid deep learning model that combines YOLOv8 and CNN to deliver high-precision and robust image classification. YOLOv8 is employed for initial object detection, and a CNN is used for fine-grained classification of detected regions. The integration of YOLOv8 and CNN is achieved through a majority voting mechanism, allowing the system to capitalize on the strengths of both models. The proposed framework has been evaluated on a custom dataset, achieving an accuracy of 85%, with improvements noted in precision, recall, and F1-score when compared to standalone models. This paper provides a detailed explanation of the preprocessing techniques, architecture design, and experimental results to validate the efficiency of the hybrid approach in real-world applications.

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

Hybrid Deep Learning Model Integrating YOLOv8 and CNN for Accurate Image Classification

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ch. Sai Sangeetha,
  • K. Bhavya Deepika,
  • Tanigundala Leelavathy

摘要

Image classification is a crucial task in computer vision, with significant applications in fields such as healthcare diagnostics, autonomous navigation, and security systems. This paper introduces a hybrid deep learning model that combines YOLOv8 and CNN to deliver high-precision and robust image classification. YOLOv8 is employed for initial object detection, and a CNN is used for fine-grained classification of detected regions. The integration of YOLOv8 and CNN is achieved through a majority voting mechanism, allowing the system to capitalize on the strengths of both models. The proposed framework has been evaluated on a custom dataset, achieving an accuracy of 85%, with improvements noted in precision, recall, and F1-score when compared to standalone models. This paper provides a detailed explanation of the preprocessing techniques, architecture design, and experimental results to validate the efficiency of the hybrid approach in real-world applications.