Klasifikasi Citra Ikan Menggunakan Convolutional Neural Network

Abstract

Fish Image Classification Using Convolutional Neural Network is an application that helping classification process. The application is used by user for knowing an information about what is the name of fish species from image that visitor captured. The application is designed by Python programming language. Method of designing the application using System Development Life Cycle. The method used in training model is Convolutional Neural Network, Data used in training process are the species data set is more than 10 species and each species is more than 1000 images, The data collected has been divided into training data and test data. In training process, Fish Image Classifiation Program produce a train loss value of 0.189203, validation loss value of 0.033459 and accuracy value of 0.991029. The evaluation process is carried out using a Confusion Matrix where the diagonal data is the correct prediction data, while the other data is the wrong prediction. By evaluating the Confusion Matrix, predicted accuracy reaches 99.1%precision and recall is 0.98. The resulting accuracy is very good accuracy so that it can predict the image inputted by the user accurately.

Classification of Sign Language in Real Time Using Convolutional Neural Network

“…The enhanced library support for Python in recent years, particularly for pandas, has made it a potent choice for data analytic jobs [9]. Computer vision, when all is said and done, means to double (or, in a few cases, compensate) human vision and, customarily, has been utilised as a part of performing routine monotonous undertakings, for example, classification in monstrous mechanical production systems [17]. Today, scrutinise on machine vision is spreading gigantically, so it is very nearly impossible to organise every last bit of its subtopics.…”