eduzhai > Applied Sciences > Engineering >

DETCID Detection of Elongated Touching Cells with Inhomogeneous Illumination using a Deep Adversarial Network

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 10 pages

Abstract: Clostridioides difficile infection (C. diff) is the most common cause ofdeath due to secondary infection in hospital patients in the United States.Detection of C. diff cells in scanning electron microscopy (SEM) images is animportant task to quantify the efficacy of the under-development treatments.However, detecting C. diff cells in SEM images is a challenging problem due tothe presence of inhomogeneous illumination and occlusion. An Illuminationnormalization pre-processing step destroys the texture and adds noise to theimage. Furthermore, cells are often clustered together resulting in touchingcells and occlusion. In this paper, DETCID, a deep cell detection method usingadversarial training, specifically robust to inhomogeneous illumination andocclusion, is proposed. An adversarial network is developed to provide regionproposals and pass the proposals to a feature extraction network. Furthermore,a modified IoU metric is developed to allow the detection of touching cells invarious orientations. The results indicate that DETCID outperforms thestate-of-the-art in detection of touching cells in SEM images by at least 20percent improvement of mean average precision.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×