Dense Captioning Of Images
i-manager s Journal on Information Technology · 2021
Recurrent Neural Network (RNN) language model that generates the captions. This project requires a system making use of computer vision to both find regions and describe them in natural language. The images are passed through a Convolutional network to identify the region features. These features then form the input for the Recurrent neural network, which generates the captions for the regions encompassing the relationships between the objects.