Dual-Modal Transformer with Enhanced Inter- and Intra-Modality Interactions for Image Captioning

Image captioning is oriented towards describing an image with the best possible use of words that can provide a semantic, relatable meaning scenario inscribed. Different models be used to accomplish this arduous task depending on context and requirement what needs achieved. An encoder–decoder model which uses feature vectors as input encoder often marked one appropriate process. In proposed work, dual-modal transformer has been captures intra- inter-model interactions in simultaneous manner within attention block. The architecture quantitatively evaluated publicly available Microsoft Common Objects Context (MS COCO) dataset yielding Bilingual Evaluation Understudy (BLEU)-4 Score 85.01. efficacy Flickr 8k, 30k datasets MS COCO results for same compared analysed state-of-the-art methods. shows outperformed when conventional models, such model.