RNN WITH ADDITIONAL CONSTANT MEMORY FOR IMAGE CAPTION GENERATION TASK

Apply

RNN WITH ADDITIONAL CONSTANT MEMORY FOR IMAGE CAPTION GENERATION TASK

Journal Title: INTERNATIONAL ACADEMY JOURNAL "Web of Scholar" - Year 2017, Vol 1, Issue 4

Abstract

Analyze and generation of automated captions for images is one of the most common problems in artificial intelligence. Existing works use LSTM (Long Short-Term Memory) as recurrent neural network cell to solve this task. After training their deep neural models can generate image caption. But there is an issue, the next predicted word of the caption depends mainly on the last predicted word, rather than image content. In this paper, we present model that can automatically generate image description and is based on a recurrent neural network with modified LSTM cell that has an additional gate responsible for image features. This modification results in generation of more accurate captions. We have trained and tested our model on MSCOCO image dataset by using only images and their captions.

Authors and Affiliations

Aghasi Poghosyan, Hakob Sarukhanyan

Keywords

ПОВЕДЕНИЕ ФЕРМЕНТАТИВНОЙ АКТИВНОСТИ ПОЧВЫ ПРИ ИСПОЛЬЗОВАНИИ ВЛАГОРЕСУРСОСБЕРЕГАЮЩИХ ТЕХНОЛОГИЙ ВОЗДЕЛЫВАНИЯ ЗЕРНОВЫХ КУЛЬТУР В УСЛОВИЯХ ЧЕРНОЗЕМА СЕВЕРНОГО КАЗАХСТАН

The article analyzes the influence of minimal and zero tillage technologies - on its biological soil activity. Results for chernozems in southern Kostanay oblast are given The paper considers one of the possible approach...

ANALIS OF HISTORICAL AND PSYCHOLOGICAL PROCESS IN THE PERIOD OF «OPEN PSYCHOLOGICAL CRISIS» IN THE CONTEXT OF THE THEORY OF ACTIVITY

This article characteristic historical psychological process in the context of the period of “open psychological crisis”, according to the conceptual approach of the outstanding scientist, Sergey Leonidovich Rubinstein....

EP ID EP458047
DOI -
Views 164
Downloads 0