Exploiting Transfer Learning Ensemble for Visual Sentiment Analysis in Social Media

Chowdhury, Etu; Chy, Abu Nowshed; Chowdhury, N. M. Istiak

doi:10.1007/978-981-33-4968-1_3

Cited by 2 publications

(1 citation statement)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, Paolanti et al [25] analyzed the sentiment of social images related to cultural heritage and compared them among VGG16, ResNet, and Inception models. Recently, Chowdhury et al [26] adopted the strategy of the ensemble of transfer learning models and employed three pre-trained deep CNN models including VGG16, Xception, and Mo-bileNet. A summary of the prior research on visual sentiment analysis is shown in Table 1.…”

Section: Related Workmentioning

confidence: 99%

An Integrated Deep Learning and Belief Rule-Based Expert System for Visual Sentiment Analysis under Uncertainty

et al. 2021

Self Cite

View full text Add to dashboard Cite

Visual sentiment analysis has become more popular than textual ones in various domains for decision-making purposes. On account of this, we develop a visual sentiment analysis system, which can classify image expression. The system classifies images by taking into account six different expressions such as anger, joy, love, surprise, fear, and sadness. In our study, we propose an expert system by integrating a Deep Learning method with a Belief Rule Base (known as the BRB-DL approach) to assess an image’s overall sentiment under uncertainty. This BRB-DL approach includes both the data-driven and knowledge-driven techniques to determine the overall sentiment. Our integrated expert system outperforms the state-of-the-art methods of visual sentiment analysis with promising results. The integrated system can classify images with 86% accuracy. The system can be beneficial to understand the emotional tendency and psychological state of an individual.

show abstract

Section: Related Workmentioning

confidence: 99%