Involuntary Lung Nodules Recognition in CT Images using 3D Features Mining and Neural Systems

The use of image processing and visualization methods for volumetric CT data sets may enable radiologists to identify tiny lung nodules more easily. For instance, it has been claimed that improving tiny nodule detection involves reconstructing CT images with close inter scan spacing and interpreting images using cine rather than film-based viewing techniques. A challenge in image processing is nodule detection. The objective is to locate the locations (and shape) of particular diseased nodules in the lungs. A nodule is a tiny, rounded lung lesion or a worm-shaped lung lesion attached to the pleura (the lung boundary) that have a radio density larger than lung parenchyma.