Down syndrome (DS), also known as trisomy 21, is a genetic disease that impairs the mental and intellectual development of a foetus. The first and second trimester ultra-sonogram pictures can be manually examined for markers of Down syndrome (DS). The VGG-19 convolutional neural network-based approach is presented in this paper for the use of US foetus photos in the identification of DS. This approach provides a rapid and cost-effective diagnostic during the first trimester of pregnancy. This suggested system's detection phase involves preprocessing, feature extraction, network training, and the detection of Down syndrome (DS). In the preprocessing, the quality and size of the data is being improved. Based on the produced feature set, this approach will classify each pixel in a foetal image as normal or abnormal. This proposed technique is evaluated utilising sensitivity, specificity, accuracy, PPV, and NPV measures.