… tags play an important role in improving the captioning performance. Instead of gleaning the
object tags … novel fully VIsion Transformer based image CAPtioning model, dubbed ViTCAP, …

… not structured primarily as an image captioning dataset, each … on the down-stream task of
image captioning [42]. Transformer-… of artwork images with the goal to generate image captions …

… Therefore, we attempt to create a specialized image-text matching task within the news
image captioning task to align vision features with entity-aware textual features. Furthermore, …

… In the standard task of image captioning, we are given an image I to be described with a
natural-language sentence Y1:T . The sentence Y1:T = 1w1, w2, ..., wT l is a sequence of T …

… Inspired by the applications of CNNs in the field of NLP, we develop a framework that only
employ CNNs for image captioning. The main contributions of this paper are: 1. We propose a …

… Recently, researchers created “Twitter A11y,” a browser extension that implements six
methods of adding alt text to images without captions on Twitter [10]. Each of these features …

… of subsequent words when generating a sentence for image captioning task [9, 10]. … for
image captioning, which makes full use of the global information of the generated target captions …

… Automatic Alt Text) system that utilizes computer vision technology to identify faces, objects,
and themes in images … results generated by DCC and the image captioning model, DCC was …

… throughout the whole captioning task without considering the … captioning network to focus
on inaccurate image objects. To tackle these challenges, we propose a novel image captioning …

… We extensively evaluate our proposals on the MS-COCO image captioning dataset and
prove the effectiveness of our method when compared with the state-of-the-arts. …