Segmentation and validation of lung cancer images using artificial intelligence (AI) is a crucial area of medical imaging that can aid in early diagnosis, treatment planning, and monitoring of lung cancer patients. This research work designed a Computer Aided design model that enabled for the detection of Lung cancer with minimum execution time and least mean square error supported by SURF, GA, and FFBP-NN. The hybrid model also tested SIFT and PCA for feature vector in the preliminary phases. The PSNR after adding the white Gaussian noise to the feature vectors stood 44.16, 46.94 and 48.68 namely for PCA, SIFT, and SURF. The present research work used 75% for the training and rest for the classification. The feature vectors obtained by SURF are modified using GA. GA results into a reduced set of SURF feature vector by 22%. The optimized feature set is then passed to SVM and trained by FFBP-NN. The classification parameters are sensitivity and classification accuracy. The proposed algorithm obtains a classification accuracy of 97.89% whereas the obtained average sensitivity is 95.8%. It is then compared with previous implemented nanglia et.al (2020) algorithm and stands a growth of 1-2.5% for both the parameters. The future possibilities may include variation in the fitness function in the present work. The future research workers may also try their hand at varying neuron count for training.

#