Generating Image Captions Using Deep Neural Networks
Mehmet Ali Can Ertuǧrul, Sevinç İlhan Omurca · 2023
The irrepressible increase in the images shared on the internet continues to keep the automatic caption generation as an active research field. Generating automated image captions using both computer vision and natural language processing techniques is a challenging field of artificial intelligence. We present a study to generate Turkish captions by using TasvirEt dataset, which is a Turkish dataset based on the Flickr8K. The feature extraction task of the images is performed by using the VGG-16 and DenseNet deep learning models. The caption generation task is realized by using the LSTM model. The use of deep neural networks allows us to improve upon the results obtained in previous studies and generate more accurate and meaningful captions. The experimental results confirm the effectiveness and performance of deep learning models in this field.