Abstract

Research on Artificial Neural Networks has been widely implemented in scientific fields, especially Geography, which identifies map image patterns to distinguish geographical features, and is applied to Google Maps. Image processing, by improving image quality so that it is easily interpreted by humans or machines (computers). The technique used is the Segmentation technique. The training method uses the Levenberg-Marquardt algorithm. The results of the training process obtained epoch values ​​of 17 out of 100 iterations. The time used during the training process is 6 seconds, the performance value = 2.06; gradient value = 8.49; Mu = 1.00; check validation = 6.0; and the regression value (R) of the training process and the whole process of 0.95889 and 0.91483 respectively. The regression value shows a linear relationship between the output network and the target. Pattern identification has succeeded in creating a new image pattern identified as land area and not land. Overall, the application made successfully fulfills the desired function, namely to help people with color blindness. 


Keywords: Artificial Neural Networks, Image Processing, Segmentation Techniques, Levenberg-Marquardt