Review Paper on Development of Hybrid Image C...

Review Paper on Development of Hybrid Image Caption Generation Method using Deep Learning

Author Name : Divya A. Nandurkar, Aachal Wanjari, Riya Hinge , Sakshi Nandardhane , Shriya Naidu

ABSTRACT The main motive of this paper is to analyze different approaches and methods used for Hybrid Image captioning and generation using Deep Learning Methods and proposing our unique method and introducing additional features to the project. Project is implemented utilizing the powerful combination of ResNet 50 architecture for image feature extraction and LSTM (Long Short Term Memory) for caption generation. This project consists of the popular and widely used Flicker 8k data set which consists of 8000 images, each of them accompanied by human annotated captions. We are expanding the image caption generation methodology to support multiple languages like Hindi and Marathi, providing an enhanced user experience for a broader audience. We are also integrating a TextTo-Speech (TTS) system in our project to add voice based descriptions along with image caption generation to help visually impaired users.