Automatic cerebral aneurysm detection in multimodal angiographic images

We propose a system to automatically detect cerebral aneurysms in 3D X-ray rotational angiography (3D-RA) images, magnetic resonance angiography (MRA) images and computed tomography angiography (CTA) images. The aneurysms are found by analyzing a blob-enhancing filtered image. Our method was tested on 65 angiographic data sets. The features leading to the best discrimination between false positives (FP) and aneurysms were identified. We achieved 96 % sensitivity with an average rate of 2.6 FP per data set in case of 3D-RA, 94 % sensitivity with an average rate of 8.0 FP per data set in case of MRA and 90 % sensitivity with an average rate of 28.1 FP per data set with CTA, respectively.