Artificial Neural Network Application to Analyze 3D Image Printing Using Artificial Intelligence in COVID-19
Paryati, Salah-ddine Krit · 2022
Three-dimensional image recognition is one of the activities used to check the original image. The method is focused on how to extract data from the available samples. The data extraction step was dividing the sample into several research areas. Then from each region the active pixel value is taken so that numerical data can be obtained as much as the available area. The numerical result data is then normalized by fixed compare, where each numeric data comes from the same data. To prove the method, the writer made a supporting application, namely char-cognition with a basic visual programming language. The final result of the analysis shows that in the pattern produced by this method the authenticity of the three-dimensional image can be well recognized. Manual examination is considered inefficient because it faces problems with the eye’s foresight and accuracy. Three-dimensional images are generally identical, but not the same in terms of shape and slope. Automatic three-dimensional image matching via a computer system is essential for better and more accurate identification of three-dimensional images. The artificial neural network method is used for extraction with various data in three-dimensional images. From the five of the simulation it can be concluded that the best Learn Rates by the lowest Error Level is 0.1 and the best Learn Rates based on the highest Match Level is 0.3. Otherwise, due to successful rates from this experiment not being measured by the low rate of Error Level, but from the high of Match Level rate, Simulation 2 with Learn Rates = 0.3 will be used in the next experiment.