Earticle

초록

영어

This research analyzes methods to enhance Urban Sound Classification performance by converting audio into Melspectrogram and MFCC images. Using the UrbanSound8K dataset, we compared single-representation and feature-level fusion strategies across CNN architectures. Results show that ResNet achieved the highest accuracy of 0.9594. However, a DenseNet-based fusion model proved more efficient, reaching a competitive accuracy of 0.9456 with fewer resources, demonstrating the potential for practical models that balance performance and efficiency.

목차

Abstract
I. INTRODUCTION
II. RELATED WORK
A. 2D Visual Representations for Audio
B. CNN-based Classification and Fusion
III. METHODOLOGY
A. Dataset and Preprocessing
B. Model Architecture
IV. EXPERIMENTS RESULTS
V. CONCLUSION
ACKNOWLEDGMENT
REFERENCES

저자

• Myeonghoe Lee [ Department of Computer Engineering Gachon University Seongnam-si, Republic of Korea ]
• Pankoo Kim [ Department of Computer Engineering Chosun University Gwangju, Republic of Korea ]
• Chang Choi [ Department of Computer Engineering Gachon University Seongnam-si, Republic of Korea ] Corresponding Author

참고문헌