Intelligence Embedded Image Caption Generator Using an LSTM-Based RNN Model

Authors

  • Dr. K. Uday Kumar Reddy, K.Vineeth Kumar, P. Varshitha, G. Swathi, N. Sasikala, E. Venugopal

Abstract

Humans have a predisposition for absorbing knowledge from whatever they see, whether it is alive or non-living. This occurrence inspired us to go down this path and investigate computer vision and how it might be utilised in conjunction with recurrent neural networks to produce captions from any image. Several other academics have worked on this problem and achieved considerable advancements, as seen by the present explosion in natural language processing-based applications. It is difficult to communicate a mental image; the structure and semantics of a phrase play a vital part in sentence building.The caption generation problem is addressed in this paper by employing an LSTM (Long-Short Term Memory) based RNN model and developing an architecture based on it to create efficient and appropriate captions by efficiently training the dataset.The Flicker8k dataset was used to train our model, and it performed admirably. The correctness of the model is assessed using conventional assessment criteria.

Published

2021-09-11

How to Cite

Dr. K. Uday Kumar Reddy, K.Vineeth Kumar, P. Varshitha, G. Swathi, N. Sasikala, E. Venugopal. (2021). Intelligence Embedded Image Caption Generator Using an LSTM-Based RNN Model. Drugs and Cell Therapies in Hematology, 10(3), 1125–1131. Retrieved from http://dcth.org/index.php/journal/article/view/1021

Issue

Section

Articles