A REVIEW ON AUTOMATIC IMAGE CAPTIONING GENERATION
Rachita Dubey, Rohit Miri · ShodhKosh Journal of Visual and Performing Arts · 2024
Image caption is a very popular approach through which descriptive language can be generated in natural form. It is a difficult task in the field of artificial intelligence to assess an image and then write captions that are appropriate using computer vision approaches. The motive of the paper is to review the related studies in image captioning. Numerous studies on image captioning have been conducted, however optimum precision is still needed for accurate and precise captioning. To create well-organized sentences, a system that considers both semantic and syntactic factors is necessary. It is necessary to get the things that are over the image and explain how they link to one another or to express the activity in accordance with the situation in the image in order to get a better caption. The goal of image captioning can be accomplished using a variety of machine learning techniques, and numerous studies have used CNN, RNN, DNN, LSTM, and other approaches. The majority of researchers evaluated their systems using a variety of benchmarks, including Flickr8K, Flickr30K, MSCOCO and many more. However, Flickr8K, which has 8092 images or challenges to test the system's performance, is the most used dataset.