• Version
• Download 7
• File Size 682.69 KB
• File Count 1
• Create Date 1 May 2023
• Last Updated 1 May 2023

Image Captioning based on Attention Mechanism

Raj Prabhu R, Santhosh Kumar S, Harris S M

^{Kongu Engineering College}
Perundurai,Erode

^{rajprabhur.19it@kongu.edu}

santhoshkumars.19it@kongu.edu

^{harrissm.19it@kongu.edu}

Abstract— With the development of deep learning, the aggregate of pc imaginative and prescient and natural language device has aroused great interest within the beyond few years. Image captioning is a representative of this filed, which makes the computer discover ways to use one or more sentences to understand the seen content material of a picture. The vast description technology method of excessive degree image semantics calls for now not handiest popularity of the item and the scene, however the ability of reading the dominion, the attributes and the connection among one’s devices. Though photograph captioning is a complicated and tough project, a number of researchers have done sizeable enhancements. Here we mainly describe 3 image captioning techniques the usage of the deep neural networks: CNN-RNN primarily based, CNN-CNN based framework and Attention Mechanism. Then we introduce the paintings of these 3 pinnacle techniques respectively, describe the assessment metrics and summarize the advantages and primary demanding situations.

Keywords— Image Captioning, Long Short Term Memory, Attention Mechanism, Deep learning, Natural language device