Team

Vi Ly: UIPath Process; Van Dang: Deep Learning

Inspiration

The deadly disease has reached epidemic, even endemic proportions in different parts of the world — killing around 400,000 people annually. In other areas of the world, it’s virtually nonexistent. Some areas are just particularly prone to a disease outbreak — there are certain factors that make an area more likely to be infected by malaria: High poverty levels Lack of access to proper healthcare Political instability Presence of disease transmission vectors (ex. mosquitos)

What it does

We want to create a web-based service powered that can classify accurately the skin cell with sign of Malaria.

Current diagnosing methods of this disease are tedious and time-consuming.

How we built it

  1. We first created a deep learning model from Keras package (with Tensorflow backend). The model consists of 3 convolutional layers and 2 dense layers.

  2. We intergrated the trained model into UIPath workflow so our testing process can be automated and we can test files by a large number with repeating tasks.

Challenges I ran into:

I tried to run the Python file directly on the UIPath server but there are many deep learning packages that is not yet available. Therefore, after several trials and failures implementing different UIPath workflow, I eventually arrive to a working process by directing UIPath Robot to run the Python file on terminal while exporing the results to txt file.

Accomplishments that I'm proud of

We are proud that this free web service can be used to aid in the early detection of Malaria. This can help to save the cost of healthcare and the problem of short-staffing at medical facility in many developing countries.

What I learned:

I am a newbie in using UIPath Studio so there are so many new information I gained from this projects. I find that UIPath is super easy to get started with and really useful in helping me automate the tedious and repetitive tasks so that I have more time in the interesting work of perfecting my deep learning models.

What's next for the Project:

I will also develop deep learning models for many other medical images such as blood smear, MRI or ultrasound.

Built With

  • keras
  • uipath
Share this project:

Updates