A Survey on Image Captioning Techniques using Deep Learning

Abstract: Image captioning is an emerging technology in the field of computer vision and machine learning. It has been an important research topic in the recent years as it involves understanding images and language modeling. Image captioning requires recognizing the important objects, their attributes and their interactions in an image. It also needs to generate syntactically correct and semantically accurate sentences. Traditional machine learning based methods and deep machine learning based methods can be used to achieve this. Deep learning-based techniques are better capable of handling the complexities and challenges of image captioning over traditional methods. In this survey paper, we aim to present a review of existing image captioning techniques focusing primarily on deep-learning based methods. We discuss the foundation of the techniques, their strengths, limitations and analyze their performances.

Keywords: Image captioning, novel concept, deep learning, computer vision.