Deep Learning-Based Architecture for Down Syndrome Assessment During Early Pregnancy Using Fetal Ultrasound Images
Journal Title: International Journal of Experimental Research and Review - Year 2024, Vol 38, Issue 2
Abstract
Down's syndrome, also known as trisomy 21, is a prevalent genetic condition characterized by intellectual disability and developmental delay in children. The current study has prioritized the development of precise screening techniques for trisomy 21 in the initial stages of pregnancy in order to facilitate prompt diagnosis. This study presents an innovative paradigm for categorizing sagittal views in obstetric ultrasound examinations, specifically identifying Nuchal Translucency, a crucial component within the fetal brain, during the 11th to 14th weeks of pregnancy. The suggested deep learning-based system effectively detects the presence of the essential cerebral structure known as Nuchal Translucency, hence aiding in the diagnosis of Down's syndrome. A dataset comprising more than 1100 pre-processed 2D sagittal-view ultrasound images was gathered to train, test, and validate the proposed convolutional neural network model. The model results were utilized to quantify neurotensin levels and assess the presence of Down's syndrome by image classification. The performance of the model was assessed by measuring its sensitivity, specificity, and area under the curve metrics. These metrics were then compared to those of human experts who had received training in prenatal and ultrasound techniques. Notably, the suggested model attained an outstanding area under the curve score of 0.97. Our study suggests the most common non-invasive method for screening pregnant women for fetal abnormalities is ultrasound, which examines the unborn organs. The application of deep learning and machine learning techniques has significantly enhanced the diagnosis. In order to detect or anticipate conditions like Down syndrome, it assesses the intercranial structures of the developing embryo during the early stages of pregnancy. Developing an improved iteration of this model could serve as the foundation for an effective automated system for diagnosing Down's syndrome in its early stages within a clinical environment.
Authors and Affiliations
Aijaz Ahmad Reshi, Shabana Shafi, Irfan Qayoom, Maria Wani, Shahida Parveen, Ajaz Ahmad
Keywords
Related Articles
- EP ID EP734957
- DOI 10.52756/ijerr.2024.v38.017
- Views 36
- Downloads 0
Evaluation of Antioxidant, Anti-inflammatory and Antimicrobial Potential of Aegel marmelos Fruit Pulp Extracts against Clinical Pathogens
In India, a wide range of medicinal plants are reported. Since ancient times, these medicinal plants have been used by people for the treatment of several diseases. Herbal medicines typically have fewer side effects comp...
Transfer of heavy metals from lake water to biota: a potential threat to migratory birds of Mathura lake, West Bengal, India
The present study was conducted in the Mathura oxbow lake, West Bengal, India to assess the severity in transferring harmful heavy metals from lake water to its biota. Plenty of common and seasonal migratory birds are de...
A Retrospective Non-Randomized Case Study Analysis of Mitomycin-C Efficacy in Patients Undergoing Combined Phacotrabeculectomy at a Tertiary Care Hospital in Salem, India
Glaucoma and cataracts, which are major contributors to worldwide blindness, frequently occur together, requiring phacotrabeculectomy as the primary therapeutic approach. This study conducted a retrospective analysis of...
Feature Ranking Using Novel Consistency Measure by Normalized Standard Deviation and Proposal of Three Novel Global Features for Online Signature Verification
Signature is a behavioral biometric that evolves throughout a person’s life. Feature extraction and ranking are very important steps towards online signature verification in order to achieve high efficiency. In our case,...
Bait preference study of a day biting mosquito in a rural area of West Bengal, India
4713 Armigeres subalbatus (Conquillett) mosquitoes were caught off human and bovid baits, out of which 33.63% from human bait and 66.37% from bovid bait. In bovid baits 21.47% and 44.9% A. subalbatus were collected from...