Image Captioning Using VGG-16 Deep Learning Model
Vishal Jayaswal, Sharma Ji, Satyankar, Vanshika Singh, Yuvika Singh, Vaibhav Tiwari
Abstract
Image captioning is a method that creates captions from images. It makes use of computer vision, natural language processing, and deep learning. The field of image captioning has evolved significantly in recent times, thanks to the application of both conventional and sophisticated deep learning approaches. This study uses two well-known benchmark datasets, Flickr8k and Flickr30k, to examine the use of the VGG-16 convolutional neural network (CNN) for producing descriptive captions. A re- current neural network (RNN) is integrated to generate coherent and contextually relevant captions after the VGG-16 model is utilized for feature extraction. Human-provided references and generated captions are compared for quality using standard assessment metrics such as BLEU. The findings have applicationsin the areas of content indexing, assistive technology for the visually impaired, and enhancing user interfaces on image-centric systems. The comparison of the Flickr8k and Flickr30k datasets sheds light on the challenges posed by the different datasets and provides guidance for future image captioning research. A list of references, a synopsis of the key findings, and suggestions for future research topics are included in the paper's conclusion.