In clinical heart disease diagnosis by real-time 3D echocardiography

it is time-consuming and tedious for doctors to manually search the best views to detect the complex cardiac anatomy

when they routinely check a large amount of patients 3D echocardiographic images. Since the apical four-chamber（a4c) view is one of the most important screening planes

the purpose of this study is to detect the a4c plane in the 3D echocardiographic image automatically. Firstly

a typical a4c image is chosen as a template image. Then to find the a4c image in a 3D echocardiographic volume data

a series of cross-sections are extracted from this data to build an ultrasound image database. Via a coarse to fine retrieval

the cross-section in the database most similar to the template is taken as the a4c image in this 3D data. Tested on 28 datasets of normal children and 22 datasets of children suffering from congenital heart disease

this method achieved the accuracy rates of 96% and 86.3%

respectively. By only one template image

this method can detect the a4c image planes in the echocardiographic volumetric datasets of different subjects. With low computational complexity and simple implementation

the proposed method gives promising results for applying the auto-detection of the a4c view to clinical diagnosis

which is significant for following computer-aided diagnosis approach such as registration and measurement.