About Cov.Base

It is an undisputed fact that DATA, and thereby EVIDENCE, is an essential pre-requisite for developing any kind of actionable clinical INSIGHT in medicine today. In fact, the world of machine learning and artificial intelligence relies entirely on the free availability of HIGH-QUALITY DATASETS. Unfortunately, as the world battles the COVID-19 pandemic right now, DATA ANALYTICS, MACHINE LEARNING AND ARTIFICIAL INTELLIGENCE tools are not part of our armory simply due to lack of availability of base data. Hence, we have created Cov.Base — a repository of clinical data of COVID-19

This is a joint effort between CARING — the Centre for Advanced Research in Imaging, Neurosciences & Genomics and IGIB — the Institute for Genomics & Integrative Biology. Its backend is powered by CARPL – The CARING Analytics Platform.

Clinical Data
Clinical notes – signs, symptoms
Treatment notes and course through hospital
Lab records


Radiology images and reports
Outcomes
Any other appropriate data that may be available
Biological Samples
Blood
Nasopharyngeal Swab
Bronchoalveolar Lavage


Viral Isolates
Viral RNA/cDNA
Any other materials that may be collected – urine, stool, biopsy etc.
Method
  • 1

COVID-19 Hospitals

  • 2

Data Collection

  • 3

Data Curation

  • 4

Release to Researchers

Model Evaluation


If you are interested in getting your AI model evaluated from us, we request you to please read below instructions to share your model.



1. Kindly share zipped docker image or pull link of your docker from dockerhub.
2. Docker should be linux based and must contain all dependencies to run your model.
3. Docker should read data(images) sequentially from a repository and must generate a output CSV containing all the labels produced by your model.
4. If your model generates multiple image outputs like heatmap, pathology localization etc corresponding to a single input image,Then separate input and output folders should be created with the same name so that it can be correlated.
5. Make sure your program runs on entire data with below command
docker run --network none -v input_data_path:/input -v output_data_path:/output docker_image_name
6. Kindly share a brief description of your model about input data for your model, model classes and input image format for your model.

After successful submission of above prerequisites, an empowered committee constituted by CSIR-IGIB will evaluate your model and share the evaluation results with you.

Kindly note that we may have numerous requests for model evaluation, we will try to share results as soon as possible.However,we expect a minimum of 7 days to fulfill such requests.



In case you have any query or for detailed SOP please reach out at nitin.batheja@igib.in

Collaborators
The Team
Request Access
get in touch


get connect
  • IGIB +91 011 29879102/03
    Caring Research +91 11 49248062
  • CSIR - Institue of Genomics and Integrative Biology,
    Mathura Road, Sukhdev Vihar,Delhi, 110025
  • Caring Research - E 19, Main Road,South Extension Flyover,Ring Road, Defence Colony, Delhi, 110024
Copyright @ 2020 CARING | All rights reserved
Protected by Copyscape