International Journal of Arts and Technology (4 papers in press)

Regular Issues

• Cosmetic Packaging Design Method Integrating Shape Grammar from the Perspective of Artificial Intelligence Generated Content: a Case Study of Song Porcelain Patterns  
  by Zhenyu Li, Zhan Gao 
  Abstract: Abstract: This paper, taking the patterns of Song Dynasty porcelain as an example, explores novel approaches to cosmetic packaging design. First, a database of Song Dynasty porcelain patterns is set up. Then, through applying the derivation method based on shape grammar, innovation is carried out, during which the forms of these patterns are deconstructed and reconstructed while the cultural symbols are retained. Next, a low-rank adaptation model is trained with the Stable Diffusion model. By combining relevant functions along with the conditional control function of ControlNet, cosmetic packaging design schemes in the style of Song Dynasty porcelain are generated. Finally, the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) is used to rank and optimise these design schemes. This comprehensive approach boosts the design's creativity, diversity, scientific nature, and rationality, offering fresh ideas for cosmetic packaging design.
  Keywords: Keywords: Artificial Intelligence; Stable Diffusion; Shape Grammar; Low-Rank Adaptation model; Cosmetic Package design;.
  DOI: 10.1504/IJART.2025.10072005
   
  
• Emotional Experience in Ethnic Vocal Music Appreciation Reflects AI Interactive Technology  
  by Yan Chen 
  Abstract: This research mainly discusses the important manifestation of the emotional experience in the music appreciation of ethnic vocal music in the AI interactive technology.The system adopts a three-layer BP neural network design, the input is the feature vector of the MIDI ethnic instrument music segment, and the output vector represents the music emotion. The music feature extraction module mainly completes the reading and analysis of music fragments in MIDI format, extracts the basic features of music and abstract features such as rhythm, melody, mode, tune, and harmony, and stores the feature information in a dynamic database. In the process of music feature extraction and recognition, the system first reads the note queue and speed information queue of ethnic vocal music in MIDI format, and then analyzes and calculates the basic characteristics of music, music mode and melody characteristics, and music rhythm characteristics.
  Keywords: AI Interaction; Emotional Experience; Music Appreciation; BP Neural Network; Music Features.
  DOI: 10.1504/IJART.2025.10072119
   
  
• Opera Audio Understanding and Synthesis via Neural Network Models: from Recognition to Generation  
  by Jie Pan 
  Abstract: This study proposes a composite model for opera audio recognition and style generation. The model integrates chaotic fingerprint coding, deep neural networks, and generative networks for style transfer. The model uses a 20-bit chaotic audio fingerprint based on logistic mapping and time-frequency peaks. This technology can achieve efficient compression and robust recognition. The accuracy of method in a noisy environment is 92.3%, which is 12.5% higher than that of traditional methods. The DNN-LightGBM cascade structure effectively models features and efficiently classifies features in 19 opera categories with an accuracy of 8895%. In terms of style transfer, generative adversarial network with orthogonal style loss function separates timbre and style and reduce Mel cepstral distortion by 18.3%, from 5.24 to 4.87. In addition, spectrum-based unsupervised linear style encoder improves the robustness of the transfer by 23.6% under various accompaniment conditions. The framework has high recognition accuracy, high-quality style transfer, and strong adaptability.
  Keywords: Opera Audio; Neural Network Models; Audio Recognition; Audio Synthesis; Generative Modeling.
  DOI: 10.1504/IJART.2026.10072740
   
  
• Assessment of Deep Learning algorithms for Damage Segmentation in Indian Murals  
  by Anshul Kumar Yadav, Ronit Kunkolienker, Dhiraj Sangwan 
  Abstract: Murals on walls of the havelis in Rajasthan have become damaged, and Image Inpainting has emerged as a potent solution to the problem. Masking damage is a crucial step for inpainting algorithms to avoid learning from damaged regions. Therefore, this study focuses on damage identification, utilising prominent architectures, including U-Net and its derivatives, as well as adversarial networks. The study also explores the effect of dense conditional random field (dCRF) and major voting ensemble algorithm. The results show that using dCRF modelling and an ensemble approach improves the average structural similarity index measure (SSIM) score from 0.9672 to 0.9703 for the test dataset. Alone, dCRF improves the output of the worst-performing model, Pix2Pix (modified), by 5.02 %. The suggested method also outperforms the generic adversarial image translation networks for mural damage segmentation on the test dataset by up to 5.45% in the mIoU and 3.61% in the mean DSC score.
  Keywords: Damage annotation; Mural restoration; Convolutional neural networks; Generative adversarial networks; Ensemble learning.
  DOI: 10.1504/IJART.2025.10072772