A novel approach for automatic detection of non-small cell lung carcinoma in CT images

Lung cancer is a conceivably deadly disease brought on predominantly by ecological factors that transform genes that encodes basic cell regularities proteins. This paper investigates the early detection of lung cancer using computer aided diagnosis system which helps to improve the life term of the patient. The Existing multimodal sparse representation based classification of lung cancer related abnormalities can't deliver the ideal recognition rate because of the irregularity of feature selection. Moreover, the computational time is also high. To overcome these limitations, a novel cancer detection method namely Enhanced Image and Feature Selection based Detector (EIFSD) is developed with higher detection rate and minimal computation time. The proposed EIFSD performs well in real time as it combines optimal features from both the enhanced and original images. The experimental results validate that the proposed method improved the Image Enhancement Factor by 19% and detection overshoot by 93% compared with existing methods. Subsequently, the proposed strategy is more appropriate for the early estimation and determination of lung cancer really well. The outcomes plainly exhibit that the recognition performance fits well for early diagnosis of lung cancer and furthermore incorporates with the present medical analysis techniques.