Author : Suramya Patel and Shilpa Gite

Picture Captioning is a testing assignment that has required a lot of information as highlight extraction to accomplish elite. In this paper, we introduce a novel neural systems design that Convert Image into Sentences utilizing a half and half bidirectional LSTM with CNN Approach, taking out the requirement for most element Engineering. We picture the development of bidirectional LSTM inside states after some time and subjectively investigate how our models make an interpretation of picture to sentence. Our proposed models are assessed on subtitle era and picture sentence recovery errands from Available Dataset. Convolutional neural networks (CNN) have turned out to be mainstream in picture handling for include extraction. We show that bidirectional LSTM with CNN Approach accomplish profoundly Performance and Significantly beat late strategies on Image Captioning.

Keywords: Bi- Directional LSTM, CNN, Image Captioning, Computer Vision, Natural Language Processing, Context Awareness

Article published in International Journal of Current Engineering and Technology, Vol.7, No.6 (Nov/Dec 2017)