Application of DMSFNN-COA technique for brand image design

Area of Science:

• Computer Science
• Artificial Intelligence
• Machine Learning

Background:

• Color significantly influences user emotions and is critical for effective brand image design.
• Accurate product color classification is essential for understanding user preferences and forecasting trends.
• Existing methods for product color brand image classification face challenges in achieving high accuracy.

Purpose of the Study:

• To introduce a novel deep learning approach, BID-DMSFNN-COA, for classifying product color brand images.
• To enhance the accuracy of brand image design by effectively categorizing colors as 'Stylish' or 'Natural'.
• To address limitations in current product color trend forecasting research.

Main Methods:

• Utilized the Mnist Data Set for image data collection and pre-processing, including noise reduction with a master-slave adaptive notch filter.
• Developed and implemented the Brand Image Design using Deep Multi-Scale Fusion Neural Network optimized with Cheetah Optimization Algorithm (BID-DMSFNN-COA).
• Employed deep learning and optimization algorithms for robust image classification.

Main Results:

• The BID-DMSFNN-COA technique achieved a remarkable 99% accuracy for both 'Natural' and 'Stylish' classifications.
• Outperformed existing methods (BID-GNN, BID-ANN, BID-CNN) which showed lower accuracy rates (65%-85%).
• Demonstrated superior performance across multiple evaluation metrics, including accuracy, F-score, precision, recall, sensitivity, specificity, and ROC analysis.

Conclusions:

• The BID-DMSFNN-COA technique offers a highly effective and accurate solution for product color brand image classification.
• The proposed method significantly enhances brand image design by improving the precision of color categorization.
• This advancement provides a more reliable tool for product color trend forecasting and brand strategy development.