Skip to main content
Fig. 2 | European Radiology Experimental

Fig. 2

From: Performance of a 3D convolutional neural network in the detection of hypoperfusion at CT pulmonary angiography in patients with chronic pulmonary embolism: a feasibility study

Fig. 2

An example of computed tomography angiography volume (a) segmentations with good agreement to manual delineations (blue outline): HU threshold (≤ -850)-derived segmentation is shown in green (b), while the prediction of convolutional neural network is shown as colourmap (≥ 0.37 probability cutoff) (c). HU Hounsfield units

Back to article page