Abstract: Image segmentation plays vital role in image understanding and practical vision systems. The main objective of medical image segmentation is to extract and characterize anatomical structures with respect to some input features or expert knowledge. Fetal heart abnormalities are the leading cause of infant mortality related to birth defects. As a non invasive and low cost imaging modality, the ultrasound imaging has become an important tool for medical diagnosis especially in the prenatal care. Soft computing algorithms are developed for the segmentation of fetal heart image and to identify the abnormalities. The proposed method of fetal heart image classification comprises of three steps namely, preprocessing, segmentation and classification. The preprocessing technique should be possible by shearlets which shows better continuum theory and low computational complexity. Adaptive K-means Fuzzy C-Means (AKFCM) clustering algorithm is utilized to segment an image into k-clusters which is computationally faster, finally Support Vector Machine (SVM) algorithm classifies the segmented image which is identified and contains components of non-parametric applied statistics Neural Networks (NNs) and machine learning. The experimental results demonstrate that the proposed method which combines the shearlet, K-Means Fuzzy clustering algorithm and Support Vector Machine (SVM) techniques are better than the other conventional techniques in the preprocessing, segmentation and classification techniques, respectively.
Shobana Nageswari and K. Helen Prabha, 2017. Classification of Fetal Heart Abnormalities based on AKFCM Clustering and SVM Techniques. Asian Journal of Information Technology, 16: 200-205.