Hybrid Modality Level Challenges Faced in Multi-Modal Sentiment Analysis
International Journal of Emerging Technology and Innovative Engineering Volume 5, Issue 6, June 2019
9 Pages Posted: 1 Jul 2019
Date Written: June 13, 2019
Aspects and sentiments expressed by human beings through behavior is computed systematically by neural networks and this approach have been tremendously appreciated throughout the research scientists nowadays. Multi modal sentiment analysis is the currently appreciated work in the deep learning field. Diverse modalities among image, videos, text, audio or fusion of the modalities here are through which human beings express their desirable information. Physically handicapped or the insecurity forces in the border of the nation can be the victims prone to be under vigilance monitoring for identifying the sentiment of actions they express. The paper provides the data sets available for various application areas under implementation and algorithms proposed till date to solve the application problems under implementation and followed by scientific challenges faced during fusion of two or modalities expressed here. The paper also includes the discussion about the implementation criteria for constructing the deep learning networks for various fusion levels of modality.
Keywords: multimodal emotion recognition system, kernel ELM, ANP, Sentiment Analysis, MKL
Suggested Citation: Suggested Citation