A 3D multi-scale Block LBP Filter for lung nodule enhancement based on the CT images

Abstract

The computer-aided diagnostic (CAD) system can greatly influence the early detection of lung cancer in radiographs and computed tomography (CT) images. And the automatic nodule detection plays an important role in the CAD systems. This paper proposed a new enhancement filter (3D multi-scale Block LBP Filter) to detect the nodules in the lung regions. By taking advantages of the nodule regions' pixel value characteristics and shape features, the proposed filter is able to enhance the nodule-like regions and restrain the line-like regions and edge regions. Moreover, Integral Image Technique (IIT) is applied accelerate the processing speed of the filter. Experimental results showed that the proposed filter can improve the accuracy and reduce the false positive.