Efficient Tumor Classification using GoogleNet Approach to Increase Accuracy in Comparison with ResNet

This paper aims to improve the accuracy in novel Tumor classification using the Google Net classifier in comparison with the ResNet classifier. Materials and Methods: Mammographic Image Analysis Society (MIAS) dataset has been collected from kaggle.com which is a repository for our study. For a sample size of 22, taken for each group, the proposed GoogLeNet classifier is compared to the existing ResNet classifier. The calculated sample size is 44. The dataset contains 5 attributes which are considered as input attributes to the classifiers. The pretest power simulation is 80% with a significance of 0.001 (p<0.05). Results: The GoogLeNet achieved an accuracy of 64.95% compared to 56.41% by the ResNet classifier. The significance value achieved is 0.001. Conclusion: From the simulation results, it is found that GoogLeNet classifier has significantly better accuracy than ResNet classifier.