Bioinformatics Thesis Proyect

Malaria(Plasmodium vivax) detection and masks generation

Create enviroment

conda create -n malaria_bio python=3 numpy imageio scipy matplotlib pandas seaborn jupyter notebook

Deep Learning Models

U-net Results

YOLO Results

Acknowledgements

We used image set BBBC041v1, available from the Broad Bioimage Benchmark Collection [Ljosa et al., Nature Methods, 2012]

• Download link | website
  Data: 1364 images (~80,000 cells)
  Hung J, Goodman A, Lopes S, Rangel G, Ravel D, Costa F, Duraisingh M, Marti M, Carpenter A. Applying Faster R-CNN for Object Detection on Malaria Images., arxiv, 2018.
  The authors want you to cite this other paper

• Thanks to Raoul for sharing U-net kernel in tensorflow https://www.kaggle.com/raoulma https://www.kaggle.com/raoulma/nuclei-dsb-2018-tensorflow-u-net-score-0-352/