Intracardiac mass detection and classification using double convolutional neural network classifier

• Published in: Maǧallaẗ al-abḥath al-handasiyyaẗ Vol. 11; no. 2 A; pp. 272 - 280
• Main Authors: Manikandan, A., PonniBala, M.
• Format: Journal Article
• Language: English
• Published: Kuwait Kuwait University, Academic Publication Council 01.06.2023
• ISSN: 2307-1877; 2307-1885
• DOI: 10.36909/jer.12237

Identification and classification of intracardiac masses in echocardiogram is one of the significant processes in the diagnosis ofcardiovascular disease. Initially, the cropping over the specific region is done in order to make the definition of the mass area. Later on, as the second step the processing of globally unique denoising technique is being implied for the removal of speckle and in order to make the preservation of anatomical structured component in the image. This is defined in terms of preprocessing and it is carried out by Patch-based sparse representation. Subsequently the description of the mass contour and its interconnected wall of the artery are being done by the segmentation mechanism denoted as Linear Iterative Vessel Segmentation model. As the prefinal stage, the processing of boundary, texture and the motion features are being carried out through the processing by double convolutional neural network (DCNN) classifier in order to determine the classification of two different masses. Totally 108 cardiac masses images are being collected for accessing the effectiveness of the classifier. It is also realized with the various state of the art classifiers as projected the demonstration of the greatest performance that has been disclosed with an achievement of 98.98% of accuracy, 98.89% of sensitivity and 99.16% of specificity that has been resulted for DCNN classifier. It determines the explication that the proposed method is capable of performing the classification of intracardiac thrombi and tumors in the echocardiography and ensures for potentially assisting the medical doctors who are in the clinical practice.