IMPLEMENTASI METODE CONVOLUTIONAL NEURAL NETWORK DALAM PENGENALAN CITRA FASHION DENGAN PENAMBAHAN DATA AUGMENTASI

The fashion industry has undergone significant transformation over the years, making it increasingly challenging for individuals to compare and find the fashion styles they desire. To simplify this process, different styles of clothing are tried on to get the right and desired look. Therefore, researchers chose to apply the Convolutional Neural Network (CNN) technique for mode classification. This approach represents one of the methodologies used to utilize computer technology in recognizing and categorizing fashion items. The main objective of this research is to assess the effectiveness of the Convolutional Neural Network method in classifying fashion using the Fashion-MNIST dataset by adding data augmentation. The dataset consists of information on various types of clothing and accessories, categorized into ten classes, including ankle boots, bags, coats, dresses, pullovers, sandals, shirts, sneakers, T-shirts and trousers. The new classification method showed improved performance on the test dataset, achieving an accuracy rate of 95.92%. In addition, this research uses an image data generator method to enhance the Fashion MNIST images, thereby preventing excessive emphasis on specific details and improving the accuracy of the results.