Abstract
In this work, we have introduced Gaussian Smoothen Semantic Features (GSSF)for Better Semantic Selection for Indian regional language-based imagecaptioning and introduced a procedure where we used the existing translationand English crowd-sourced sentences for training. We have shown that thisarchitecture is a promising alternative source, where there is a crunch inresources. Our main contribution of this work is the development of deeplearning architectures for the Bengali language (is the fifth widely spokenlanguage in the world) with a completely different grammar and languageattributes. We have shown that these are working well for complex applicationslike language generation from image contexts and can diversify therepresentation through introducing constraints, more extensive features, andunique feature spaces. We also established that we could achieve absoluteprecision and diversity when we use smoothened semantic tensor with thetraditional LSTM and feature decomposition networks. With better learningarchitecture, we succeeded in establishing an automated algorithm andassessment procedure that can help in the evaluation of competent applicationswithout the requirement for expertise and human intervention.