A study on Down syndrome detection based on Artificial Neural Network in Ultra sonogram images

Down syndrome is a genetic disorder in which disrupts infants, physical and cognitive development. Down syndrome is characterized by the absence of nasal bone during the late first trimester of pregnancy. Presently Down syndrome is identified by visually examining the ultra sonogram image of foetus of 11 to 13 weeks of gestation for the presence of nasal bone. The visually identification by the change in the contrast of nasal bone region of ultra sonogram is a very difficult task. So the image processing based features extraction by considering various parameters have been extremely important. This paper provides comprehensive survey on various medical imaging techniques that can be effectively used for detecting the syndrome in the early stage of pregnancy. Our proposed survey consider different methods based on the various parameters extracted using a series of operations such as Region Of Interest (ROI), Nasal Bone (NB) segmentation using morphological, Otsu thresholding and logical operations from the ultra sonogram images, both in spatial domain as well as transform domain using Discrete Cosine Transform (DCT) and wavelet transforms. The extracted data is normalized and used to train classifiers like Back Propagation Neural Network (BPNN). This paper illustrates overview of various states of methods available in the Down syndrome detection and comparison analysis of each method is discussed.