In medical volume visualization, one of the main goals is to reveal clinically relevant details from the CT study by classification of the data, i.e. the coronary arteries, without obscuring them with less significant parts. Usually, the classification is carried out by defining multi-dimensional transfer functions which assign specific visual attributes to the voxels which express the features of interest. Unfortunately, this can become a fairly complex task, generally accomplished by trial and error even for the experienced user. Many sophisticated semi-automatic and automatic approaches for volume classification have been published in the past, which rely either on the overall quality of the rendered image or on a general boundary detection between different materials rather than on an insight as to what makes the transfer function appropriate for a specific feature in the dataset. This paper presents an efficient way for automatic transfer function generation based on neural networks. We describe how to use neural networks to detect distinctive features of the volume data and how this information can be used to provide the user with a semantic view on the automatic data classification.
Nema pronađenih rezultata, molimo da izmjenite uslove pretrage i pokušate ponovo!
Ova stranica koristi kolačiće da bi vam pružila najbolje iskustvo
Saznaj više