Traffic Image Captioning in Bahasa Indonesia Using Convolutional Neural Network and Long Short Term Memory

Isma Artyani,Fenty Eka Muzayyana Agustin,Victor Amrizal,Yusuf Durachman

Intelligent Transport System is one of adjustments of the artificial intelligence technology which is used to monitor the traffic on the road. The system uses image or video as media of information dissemination to the society. One of the ways to get the information from the image is read the image caption. Image caption helps us to understand implicit information from the image. For this reason, it is necessary to put caption on the traffic image in order to ease the society getting the information from Intelligent Transport System. In creating the image caption, computer requires a combine method, Computer Vision to process the image and Natural Language Processing to process image caption text. In this research, simulation Convolutional Neural Network and Long Short Term Memory method are used to produce a model, which will be used for generate image captioning on traffic image. This study produced two models. Model-1 used VGG-16 for extracting image feature and Model-2 used ResNet50 for extracting image feature with the score BLEU {1,2,3,4} from each models in amount of {0.66 , 0.50, 0.43, 0.30} on Model-1 and {0.67, 0.51, 0.42, 0.27} on Model-2.

Volume 12 | Issue 6

Pages: 2524-2535

DOI: 10.5373/JARDCS/V12I6/S20201212