Image Caption Generation using Deep Learning Techniques

With use of natural language processing(NLP) and computer vision(CV) the problem of identification of contents of an image is solvable using artificial intelligence. In this paper, a generative model is presented which uses a deep recurrent architecture that combines recent advances in machine translation and CV which is used to generate natural sentences describing an image. An image caption generative model is created using Convolutional Neural Networks (CNN) and Long short term memory (LSTM) which uses different techniques of CV to understand an image's context and describe it in a manner native to English. For training purpose, the Flicker8K [17] dataset is used. It consists of 8000 unique images. Caption generation is one of the interesting areas of Artificial Intelligence that has many challenges. After generating the caption, it is converted into audio format. The result shows that the designed system is providing promising results w.r.t caption generation