Malaria Edge Detection with OpenVINO Toolkit

Using OpenVINO optimized image classification fast.ai model to learn and classify healthy and infected blood smear malaria images.

Data set

Malaria Datasets from the NIH — a repository of segmented cells from the thin blood smear slide images (blood films) from the Malaria Screener research activity

Why this Project

Every 2 minutes, a child dies of malaria link and according to the World Health Organization's GHO data,globally, an estimated 3.4 billion people in 92 countries are at risk of being infected with malaria and developing disease (map), and 1.1 billion are at high risk (>1 in 1000 chance of getting malaria in a year). According to the World Malaria Report 2018, there were 219 million cases of malaria globally in 2017 (uncertainty range 203–262 million) and 435 000 malaria deaths.But early diagnostics and treatment of malaria can save lives and prevent casualties.This is the reason of developing this project to make malaria detection and diagnosis fast, easy, and effective using Intel's OpenVINO based EDGE AI Model.

How to use this Project

Step 1: Open the notebook.ipynb (Available in this repository).

Step 2: Click on this Button

Step 3: Hover to the Menu bar --> Click on File Button --> Click on Save a copy in Drive

Step 4: Run the First Cell i.e. Mounting the drive

Step 5: Hover to the Menu bar --> Click on Runtime --> Run after

That's it, and it will do all the work on its own and soon, you will see how the inference will be applied on the input images (blood films) and result in AI based detection of whether the input blood film is of a Parasitized Patient or an Uninfected Person.

Flow

Tutorial Video

Reference:

https://colab.research.google.com/drive/1TdjV6bSrgSAL6RcGsYzCaWMdaqaqCYID https://towardsdatascience.com/deep-learning-and-medical-image-analysis-for-malaria-detection-with-fastai-c8f08560262f