ISOO_DL: Instance segmentation of overlapping biological objects using deep learning

Anton Böhm, A. Ücker, T. Jäger, Olaf Ronneberger, Thorsten Falk

15th International Symposium on Biomedical Imaging (ISBI), IEEE: 1225--1229, 2018

Abstract: Image segmentation is an important first step for the quantitative analysis of biomedical images. We present a method to simultaneously segment and classify translucent overlapping objects in 2D images. For this we propose an approach using a fully-convolutional neural network simultaneously solving two tasks: object detection and instance segmentation. Object detection predicts reference points, object class labels and sizes. To solve the problem of multiple labels per location, we lift our label-space from 2D to 3D, resulting in a non-overlapping representation of the instance masks. To our knowledge it is the first method that handles overlapping biological objects using deep learning making it easily applicable to a large variety of challenging datasets.

Paper Publisher's link

Images and movies

BibTex reference

@InProceedings{BRF18a,
  author       = "A. B{\"o}hm and A. {\"U}cker and T. J{\"a}ger and O. Ronneberger and T. Falk",
  title        = "ISOO_DL: Instance segmentation of overlapping biological objects using deep learning",
  booktitle    = "15th International Symposium on Biomedical Imaging (ISBI)",
  pages        = "1225--1229",
  month        = " ",
  year         = "2018",
  publisher    = "IEEE",
  keywords     = "convolution;feedforward neural nets;image classification;image representation;image segmentation;learning (artificial intelligence);object detection;instance segmentation;biological objects;deep learning;image segmentation;quantitative analysis;biomedical",
  url          = "http://lmbweb.informatik.uni-freiburg.de/Publications/2018/BRF18a"
}

Other publications in the database