Combination of volume‐rendering 3D surface modeling and medical illustration to capture the living fetus

Abstract

Objective A good medical illustration renders essential aspects of a procedure or condition faithfully, yet idealizes it enough to make it widely applicable. Unfortunately, the live fetus is generally hidden from sight, and illustrating it relies either on autopsy material or manipulated newborn images. High‐definition volume rendering of diagnostic imaging data can represent hidden conditions with an almost lifelike realism but is limited by the resolution and artifacts of the data capture. We have combined both approaches to enhance the accuracy and didactic value of illustrations of fetal conditions. Methods Three examples, of increasing complexity, are presented to demonstrate the creation of medical illustrations of the fetus based on semiautomatic computerized posthoc manipulation of diagnostic images. Results The end product utilizes the diagnostic accuracy of ultrasound and magnetic resonance imaging of the fetuses and the spatial manipulation of 3D models to create a lifelike, accurate and informative image of the fetal anomalies. Conclusion Volume‐rendering and 3D surface modeling can be combined with medical illustration to create realistic and informative images of the developing fetus, using a level of detail that is tailored to the intended audience.

Image rendering efficiency improvement based on deep autoencoder in virtual environment

“…Different from image-based modelling, image rendering does not need to use multiple two-dimensional images to build a three-dimensional geometric model of an object, and then re-project the built model into the original scene to realize visual observation from different angles. Image rendering technology can be roughly divided into two methods according to whether geometric information is needed or not [8,9]. One method is to build a rough 3D model scene according to the existing geometric information.…”