As a result of various studies on data analysis on images with limited human power, SIFT and
K-Nearest Neighbor (KNN) algorithms have been proposed as feature extraction methods for image
classification in image data. The method is designed by a person, and it is difficult to analyze all the
complex patterns of the image data.
In this paper, we extend the convolutional neural network (CNN) through deep learning to image data
to learn the topological structure of the image and classify the image by using minimum preprocessor and
convolution neural network The parameter setting for finding the fastest optimal point of the image improves the accuracy in image classification and reduces the loss rate.
Also, in the evaluation of the accuracy of the training data and the evaluation data, a study was
conducted to prevent the overfitting by reducing the difference between the two values. At the same time,
we study the parameter setting and model instantiation for improvement of image data classification by
comparing image classification through convolution neural network and accuracy and loss rate in image
classification through classifiers such as AdaBoostClassifier and RandomForestClassifier.