It is difficult to identify lung cancer by the radiologist from the computed tomography scan. A Novel CAD system based on hierarchical scale vector quantization scheme is proposed in this paper. The high level VQ gives accurate lung segmentation from the chest volume compared to the commonly used simple thresholding approach. The low level VQ is used for identifying Initial nodule of the lung. Its effective and computationally efficient. Rule based filtering operation is combined with feature based SVM technique. The proposed system was conducted and they having at least one juxta pleural nodule .The experimental result says that our CAD system obtained 82.7% of overall sensitivity at a specificity of four FP’s Per scan and 89.2% sensitivity at 4.14 FP’s per scan for the preformation of Juxta pleural nodule compared to CAD System. The proposed system shows a good performance for its fast and adoptive detection of Pulmonary Nodule via CT imaging.
