New Delhi, 24 September (IANS). Researchers from Indian Institute of Technology (IIT) and All India Institute of Medical Sciences (AIIMS) Jodhpur have used AI for better identification of malnutrition in children.
Information about this new method has been given in Open-access Journal MICCAI. The problem of malnourished children is one of the most serious health challenges globally and this method accurately assesses it.
The study dynamically adjusts the workload using domainedpta (a new multi-functioning teaching framework) using knowledge and mutual information.
This can make this system more accurately estimate the child’s height, weight and the major anthropometry (human) measurements such as the mid-upper arm perimeter (MUAC), as well as classification of malnutrition conditions such as dwarfness, weakness and low weight.
Although these measurements are also evaluated using traditional test methods, they still present various challenges to the user. It takes time to measure each aspect one by one.
“We can estimate the nutritional status without the need for complex and time-taking anthropometric measurements in this method,” examples Khan, who is doing a doctorate with medical technology at IIT-AIMS, who leads the study.
Khan further said, “This will especially test malnutrition in areas with limited resources, fast, more accessible and highly measured.”
The cornerstone of this research is anthrotion. It is a unique dataset of its kind which includes 16,938 multi-pose photos of 2,141 children, which have been collected from both the Clinical (AIIMS Jodhpur) and the Community (Government School in Rajasthan).
Through rigorous experiments, domainedpet has made significant improvements compared to existing multitask learning methods and provided a reliable AI-managed solution to accelerate malnutrition detection worldwide.
“This research is an important step towards similar healthcare access.”
He added, “By combining AI and domain expertise, we can empower healthcare personnel and public health systems with equipment that are affordable, accurate and measured.”
-IANS
KR/
