KLASIFIKASI CITRA BURUNG CENDRAWASIH MENGGUNAKAN METODE CONVOLUTIONAL NEURAL NETWORK (CNN)

This research aims to classify images of birds of paradise using the Convolutional Neural Network (CNN) method with the VGG-19 architecture. Birds of paradise, with their beautiful feathers and interesting variety of species, are the subject of this research to increase identification accuracy through AI technology. The dataset used consists of 1785 images covering five different types of birds of paradise, obtained from the internet. The classification process is carried out by applying data augmentation techniques to overcome overfitting and underfitting problems.
The research method involves several stages, including dataset collection, preprocessing, dataset sharing, and implementation of the CNN model. Testing was carried out in two scenarios: without augmentation and batch normalization, and with augmentation and batch normalization. Test results show that the use of data augmentation and batch normalization significantly increases model accuracy from 85% to 97%.
Thus, this research not only confirms the effectiveness of the CNN method with the VGG-19 architecture in classifying images of birds of paradise, but also makes an important contribution to conservation efforts for birds of paradise through the development of an automatic identification system. It is hoped that this research can become the basis for further research in the field of conservation and image recognition based on AI technology.