The impact of music visualization model by using internet of things techniques and deep neural network

The allure of music lies in its ability to engage human auditory and visual perception. Therefore, researching music is a significant endeavor. This study aims to explore an innovative music visualization model that integrates deep neural network technology, focusing on the field of the Internet of Things (IoT) and human-computer interaction. The objective is to assist in identifying musical emotions and, by introducing the Convolutional Neural Network (CNN) and IoT technology, design an efficient music visualization analysis model. Through experimental testing on a dataset obtained from NetEase Cloud Music, this study demonstrates the advantages of this CNN-based music visualization model in terms of accuracy, binary classification capability, complexity, and generalization ability. The proposed model surpasses other algorithms in terms of visual fidelity, structural similarity (SSIM) index, and peak signal-to-noise ratio (PSNR), showcasing effective data augmentation capabilities. While the model exhibits a certain level of complexity and slightly longer processing times, results with an SSIM score of 0.712 and a PSNR score of 16.893 highlight its strong generalization ability. The study reveals the significant practical value of CNN-based music visualization analysis models in the analysis and visualization of music emotion data. However, it is important to note that the dataset used in this study is limited to NetEase Cloud Music and may carry some bias, not fully representing the entire music domain or characteristics of other music streaming platforms. Therefore, the study suggests incorporating diverse music datasets to enhance the model's generalization ability and applicability. In conclusion, CNN-based music visualization analysis models possess robust generalization capabilities and application potential. They hold significant value and insights for academic research and industrial applications in music-related fields.