Unifying Vision and Language: A Versatile Approach to Image Captioning
Bhanu Pratap, Aakash Bhardwaj, Amita Goel, Ms. Nidhi Sengar -, Vasudha Bahl · International Journal of Research Publication and Reviews · 2024
Significant strides in image captioning, a pivotal domain in AI, aim to emulate human-like comprehension of visual content.This study introduces an innovative methodology integrating attention mechanisms and object features into an image captioning framework.The research leverages the Flickr8k dataset to explore the synergistic fusion of these elements, enhanc-ing both image understanding and caption generation.Addition-ally, this investigation showcases the practical application of this model through a user-centric interface employing FASTAPI and ReactJS, featuring multilingual text-to-speech translation.The empirical findings underscore the effectiveness of this approach in propelling the advancement of image captioning technology.The tutorial provides a comprehensive blueprint for constructing an image caption generator, employing a Convolutional Neural Network (CNN) for precise image feature extraction and a Long Short-Term Memory Network (LSTM) for Natural Language Processing (NLP).