Small Sample Object Detection Based on Improved YOLOv5

DOI

Description

Object detection is widely used in various production and life, such as mask detection and recognition during the epidemic, face recognition with masks. Object detection algorithm based on deep learning has always been an important research content and implementation method in the field of object detection. Due to the large number of lead seals and fuses, their locations are not fixed, the lead seals and fuses have difficulties such as few sample datasets, complex target background and easy to be blocked, and strong reflective interference, and the conventional image processing methods are difficult to solve the problem of effective object recognition. In this study, by expanding the datasets, using different data enhancement methods, and training in the improved algorithm, the detection accuracy, detection speed, and adaptability were effectively improved.

Journal

Keywords

Details 詳細情報について

  • CRID
    1390018506585197312
  • DOI
    10.5954/icarob.2024.os24-3
  • ISSN
    21887829
  • Text Lang
    en
  • Data Source
    • JaLC
    • Crossref
  • Abstract License Flag
    Disallowed

Report a problem

Back to top