New Delhi, July 25 (IANS) Scientists at the Institute of Advanced Study in Science and Technology (IASST), an autonomous institute of the Department of Science and Technology (DST) on Thursday announced the development of a cutting-edge computational model that significantly enhances the diagnosis of cervical dysplasia.
The condition involves the abnormal growth of cells on the surface of the cervix and can be a precursor to cervical cancer.
The study, published in the journal Mathematics, emphasised the importance of precise pattern identification and classification in the diagnosis and management of cervical cell dysplasia.
"Our goal was to create a model that not only offers unparalleled accuracy but also operates efficiently with minimal computational resources," said Dr. Lipi B. Mahanta, from the IASST.
The team's comprehensive approach involved experimenting with various colour models, transformation techniques, feature representation schemes, and classification methods to optimise the machine learning (ML) framework they developed.
The research involved testing the model's performance using two datasets -- one sourced from healthcare centres across India and another publicly available dataset.
By employing the Non-Subsampled Contourlet Transform (NSCT) and the YCbCr colour model, which is a specific method of representing colours in images, the model achieved an impressive average accuracy rate of 98.02 per cent.
This high level of accuracy highlights the model's potential to be a transformative tool in medical diagnostics.
"Our findings could lead to significant advancements in the early detection of cervical dysplasia, providing healthcare professionals with more accurate diagnostic tools," Dr. Mahanta noted.
The breakthrough is expected to enhance diagnostic precision and improve treatment outcomes for patients at risk of cervical cancer.
The development of this model marks a significant step forward in medical technology, offering new hope for early and accurate detection of cervical dysplasia.
![]() |
![]() |