Cancer Reviewer

Project Title

Cancer Reviewer

Project Area of Specialization Artificial IntelligenceProject Summary

Cancer Reviewer is a research based on deep learning models to detect and classify cancers from Blood Samples (Whole Slide Images). It is improvement in researches in this field. We achieved accuracy of 97% on training and testing.

Project Objectives

Cancer detection is a common, critical and time-consuming task for oncologists. Mis detection can cost patients extra time, money and put patient;s life at risk. There is always possibility of human errors in decisions of oncologists about detecting cancer.

Objective of this project is to provide assistanance to oncologists to detect cancer in no time.

Project Implementation Method

Whole Slide Images (Blood Images) are converted into patches.

These patches are further processed using Stain Normalization.

Then Monte-Carlo EM Method is applied to label dataset.

Then we splitted dataset into train and test set.

Then multiple neural networks are trained on training dataset generated with Monte-Carlo EM Method.

Benefits of the Project

It will detect cancer as we achieved ground breaking results (97% Accuracy).

Detecting time of cancer is 0.2 seconds. Whereas a oncologist will take 3 weeks to detect cancer.

It will save many lives as we will save time and cost to check if a person has cancer or not.

Technical Details of Final Deliverable

A working application with trained Neural Networks at backend to detect and classify cancers in no time to help oncologists.

Final Deliverable of the Project Software SystemType of Industry IT , Medical Technologies Artificial Intelligence(AI), Big DataSustainable Development Goals Good Health and Well-Being for People, Life on LandRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	70000
GeForce RTX 2070 (Graphics Card)	Equipment	1	70000	70000