In this paper, we utilized LeNet5 classifier to detect the melanoma regions in the input image after suitable training. The contribution of the paper includes reduced time and effort by reducing the computation time to process the high-level semantic information by proper setting of the parameters. LeNet5 is an approach of a large multi-layer neural network that treated each pixel as a separate input. The strong spatial correlations between the images, it is not recommended to use single pixels as input features in the first layer. The simulation is conducted to test the efficacy of the classifier against various images from MNIST: HAM10000 dataset in a robust manner. The results show that the proposed method achieves an accuracy of 99% than the existing classifiers.