Surface Texture-Audio Correlation for MultiModal Recognition Enhancement

In this research, we study the correlation between surface textures and audio signals produced during the movement using advanced metrics such as Structural Similarity Index Measure (SSIM), Peak Signal to Noise Ratio (PSNR) and Histogram of Oriented Gradients (HOG). Results show that rough textures achieve higher SSIM values up to 0.9914, and result in greater preserved structure than the smoother textures. Rough textures reached a PSNR at 29.85 dB, indicating less noise and higher fidelity. In addition, we found the HOG similarity to be as high as 0.9517 for rough textures exhibiting consistent gradient patterns. These results show that rough textures have an advantage when stability of structure and signal quality is desired, contradicting the commonplace preference of smoother textures in multi-modal recognition systems. The results of the study highlight the robustness of integrating visual and auditory modalities for improved recognition accuracy, and have applications in texture generation, biometric authentication, multimedia processing and virtual reality. Further work will be directed towards real time system integration as well as analysis of the broader texture range.