Deep Learning Techniques in Image Description

Abstract

The current era of high end GPU with the exploring set of data requires in-depth analysis using deep learning techniques with the combination of computer vision and natural language processing has attracted huge attention of IT industry in the recent past. The domain of Image captioning is a great revolution with respect to deep neural networks. In this paper, the hybrid model is proposed to retrieve Image description from the dataset. The motivation behind this work to explore the deep learning techniques in image dataset which is quite impressive, although to implement the processing of different models was quite challenging to get the better accuracy of results. The paper uses the hybrid algorithm CNN-CNN and CNN-RNN based model to describe the images and visual experience of that images in the form of meaningful sentence.