Overview / Usage

The project underlines the process using Intel Open Vino Toolkit to do inference and recognize new Drugs ,The data sets are created using Customised version of Mobile SSD and converting it into Tensorflow Model and then recognised accordingly.

AI on healthcare
1.Utility of the model
2.What solutions we are trying to find
3.Benefits

Utility of the Model:-
Early drug discovery is essential part of pharmaceutical sector.
It incur lots of cost and time taking too. We have to feed it through high throughput screening machines.
If we work around a solution we get a resultant that helps us generate virtual molecules on demand.
More utilities are there are virtual simulators the drug generated reduces the entire optimization process as it does not have to go through the entire drug discovery pipeline.
Using Reinforcement Learning we don’t have to apply brute force method and allow for designing chemical libraries to reduce each and every process.

What we are trying to find
We are using the data science and AI methods to find a virtual molecule which helps us to take a decision and reduce time for unnecessary walkthrough for the process.

Methodology / Approach

We have used Intel Open Vino toolkit and WinML for our project.
The image datasets for drugs have been generated using customised mobile ssd using Intel Annotation toolkit and Tensorflow models being generated and from that we generate xml and bin file
From xml file we do inference process a new image of drug and from that we are able to label which class the drug is from the trained model.
WinML part
Using DEEPCHEM we generate models which are converted to WinML format from winml tools.
The process is combination Intel Open Vino + WinML

Technologies Used

Python
Intel Open Vino Toolkit
Tensorflow
WinML
Intel Annotation tool for the image dataset.

Collaborators