An Adaboost-backpropagation neural network for automated image sentiment classification

Area of Science:

• Computer Vision
• Artificial Intelligence
• Psychology

Background:

• The proliferation of digital images necessitates advanced methods for automatic emotional content analysis.
• Extracting and classifying implicit emotional semantics from images presents a significant technological challenge.

Purpose of the Study:

• To develop and evaluate an efficient method for automatic emotional semantic classification of images.
• To improve the accuracy and performance of image-based emotion recognition systems.

Main Methods:

• Proposed an emotional semantic classification method integrating Adaboost and backpropagation (BP) neural networks.
• Utilized the Ortony, Clore, and Collins emotion model to describe image emotions.
• Constructed a strong classifier by combining 15 BP neural network outputs via the Adaboost algorithm.

Main Results:

• Experiments on 600 natural scenery images demonstrated superior performance compared to traditional BP neural networks.
• Achieved an approximate 15% increase in accuracy rate compared to previously reported methods.
• The Adaboost-BP neural network model showed significant practical value in emotional image classification.

Conclusions:

• The proposed Adaboost-backpropagation neural network method offers a robust solution for automatic emotional image classification.
• This research lays the groundwork for future advancements in sentiment analysis of visual data.
• The method has demonstrated practical utility and improved efficiency in classifying emotional semantics in images.