Abstract

Neural Style Transfer (NST) has revolutionized digital art by synthesizing the content of one image with the artistic style of another image, relying fundamentally on feature extraction from pre-trained Convolutional Neural Networks (CNNs). A critical, yet often heuristically determined, factor in this process is the selection of specific CNN layers for content and style representation, which significantly affects the final aesthetic quality. This study presents a systematic experimental investigation of the effect of layer selection on the perceptual quality of stylized images. We compared the performance of the established VGG19 architecture with the more modern and efficient EfficientNet-B0, as well as a Custom-Designed CNN, across various combinations of content and style layers. The rating is based on three important criteria: color palette transfer, visibility of the artistic technique (e.g., brushstrokes), and level of detail generalization. By objectively evaluating these criteria, this study aims to provide real, non-heuristic guidance for optimizing creative synthesis. The findings offer valuable insights for researchers and practitioners, enabling the informed selection of network architectures and layer configurations to achieve superior and predictable artistic outcomes in Neural Style Transfer applications.

References

• Alexandru, I. C., Nicula, C., Prodan, C., Rotaru, R.-P., Voncila, M.-L., Tarba, N., & Boiang (2022). Image Style Transfer via Multi-Style Geometry Warping. Applied Sciences, 12(12), 6055. https://doi.org/10.3390/app12126055
• Artistic Arbitrary Style Transfer. (2022). https://doi.org/10.48550/arxiv.2212.11376
• Galerne B., Raad L., Lezama J., Morel J. (2022). Scaling Painting Style Transfer. arXiv.Org, abs/2212.13459. https://doi.org/10.48550/arXiv.2212.13459
• Gatys , L. A., Ecker , A. S., & Bethge , M. (2015). A Neural Algorithm of Artistic Style.arXiv preprint arXiv :1508.06576 .
• Gatys, L. A., Ecker, A. S., & Bethge, M. (2016). Neural Algorithm of Artistic Style. Journal of Vision, 16(12), 326.
• Hinton, G. E., & Salakhutdinov, R. R. (2006). Reducing the dimensionality of data using neural networks. Science , 313(5786), 504–507.
• Johnson, J., Alahi, A., & Fei-Fei, L. (2016). Perceptual Losses for Real-Time Style Transfer and Super-Resolution . European Conference on Computer Vision (ECCV).
• Kashyap , K. H., Garg , M., Fargose , S., & Nair, S. (2025). Dynamic Neural Style Transfer for Artistic Image Generation using VGG19. https://doi.org/10.48550/arxiv.2501.09420
• LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning was applied to document recognition. Proceedings of the IEEE, 86(11), 2278–2324.
• McCulloch, W. S., & Pitts, W. (1943). A Logical Calculus of Ideas Immanent in Nervous Activity. Bulletin of Mathematical Biophysics, 5(4), 115–133.
• Md Redoan Hosen, Md Borhan Hosen, Afzalur Rahaman , Yasin Arafat and Abhijit Pathak (2023); ENVIRONMENTAL POLLUTION ANALYSIS AND PREDICTION OF INFLUENTIAL FACTORS: A DATA-DRIVEN INVESTIGATION Int. J. of Adv. Res. 11 (Oct). [ 323-336] (ISSN 2320-5407).
• Miah , J., Cao, D. M., Sayed , M. A., & Haque , M. S. (2023). Generative AI Model for Artistic Style Transfer Using Convolutional Neural Networks. Journal of Computer Science and Technology Studies, 5(4), 78–85. https://doi.org/10.32996/jcsts.2023.5.4.9
• Mirza Maria Moon, Sadia Afrin, Abhijit Pathak . Classification of User Reviews on Online Travel Booking Applications in Bangladesh using Multinomial Naïve Bayes. International Journal of Computer Applications. 187, 37 ( Sep 2025), 56-61. DOI=10.5120/ijca2025925646
• Rosenblatt, F. (1958). The Perceptron: A Probabilistic Model for Information Storage and Organization in the Brain. Psychological Review, 65(6), 386–408.
• Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986). Learning representations by backpropagating errors. Nature , 323(6088), 533–536.
• Ruta , D., Gilbert, A., Motiian , S., Faieta , B., Lin, Z. L., & Collomosse , J. (2022). HyperNST : Hyper-Networks for Neural Style Transfer. 201–217. https://doi.org/10.48550/arXiv.2208.04807
• Saiful Kabir, Sihabul Islam Safin, Marjahan Tanjin, Himu Akter, Rajib Ghose, Abhijit Pathak . Predicting Loan Repayment Reliability in Cooperative Societies using Naïve Bayes Classifier: A Data Mining Approach for Risk Mitigation and Decision Support. Int Journal C.omput. Appl . 186, 36 ( Aug 2024), 16-23. DOI=10.5120/ijca2024923937
• Sethi, P., Bhardwaj, R., Sharma, N., Sharma, D. K., 38 , Srivastava, G. (2024). A deep learning-based neural style transfer optimization approach.Intelligent Data Analysis, 1–15. https://doi.org/10.3233/ida-230765
• Seyed , S. H., Cansever , A., & Hart, D. (2025). Improving Masked Style Transfer using Blended Partial Convolution. arXiv.Org, abs/2508.05769. https://doi.org/10.48550/arxiv.2508.05769
• Simonyan, K., & Zisserman, A. (2014). Very Deep Convolutional Networks for Large-Scale Image Recognition. arXiv preprint arXiv:1409.1556.
• Tan, M., & Le, Q. (2019). EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks. International Conference on Machine Learning (ICML).
• Werbos, P. J. (1974).Beyond Regression: New Tools for Prediction and Analysis in the Behavioral Sciences. Ph.D. Thesis, Harvard University Cambridge.
• Xian‐Fang Li, Han Cao, Zhaoyang Zhang, Jiacheng Hu, Yuhui Jin, Zihao Zhao -12 Nov 2024