- A team from the Indian Institute of Technology, Madras (IIT) has done the unfathomable by using deep learning AI techniques to diagnose cancer accurately and quickly
- The histopathology tissue analysis is the most common way to identify cancer and these scientists may have just changed the game by making WSI detection easier
- With this new research, IIT-M researchers have overcome multiple technical challenges, the study published in Nature said
Detecting cancer with artificial intelligence? A team from the Indian Institute of Technology, Madras (IIT) has done the unfathomable by using deep learning AI techniques to diagnose cancer accurately and quickly.
The team, comprising Dr. Mahendra Khened, Avinash Kori, Haran Rajkumar, Prof. Ganapathy Krishnamurthi, Prof. Balaji Srinivasan from different engineering departments at IIT-M developed a deep learning mechanism using Whole Slide Imaging (WSI) which helps in detecting different kinds of cancers.
The histopathology tissue analysis is the most common way to identify cancer and these scientists may have just changed the game by making WSI detection easier. According to the study, WSI is becoming more common to detect cancers because the method shows high resolution images.
But reading and processing WSI isn’t simple and requires tedious work – regardless of who is undertaking it, be it computational algorithms or human beings.
With this new research, IIT-M researchers have overcome multiple technical challenges, the study published in Nature said. The impediment posed by the large size of WSI images was overcome by IIT scientists using automated analysis which tackled images on patch-by-patch basis.
The other problem with diagnosing cancer using AI deep learning is insufficient training data for there are simply not enough datasets available for AI to learn how to make its diagnosis. To overcome this drawback, the scientists developed special training methods.
In addition, there is variation in different WSI images obtained from different labs. The IIT-M scientists developed preprocessing techniques to avoid inconsistency.
The deep learning AI model was able to successfully detect various forms of cancer including breast cancer, liver cancer and colon cancer. The challenges in terms of detection were effectively overcome in all types of cancer by this framework, and it was able to detect different forms of cancer with great accuracy.
Do you think such researches could change the healthcare industry and improve the life of millions of people that suffer serious illnesses in India?
Original News Link
https://www.indiatimes.com/technology/news/iit-madras-scientists-cancer-detection-ai-framework-548918.html
