Article: Deep learning-based innovative product design driven by social network data Journal: International Journal of Information and Communication Technology (IJICT) 2026 Vol.27 No.49 pp.59 - 78 Abstract: Social network data has become a vital resource driving product innovation and design. Current research struggles to fully uncover users' emotional needs toward products when dealing with unstructured, high-dimensional social data, resulting in subpar product quality. To address this, this paper first employs a multi-scale attention network to analyse product emotional needs, capturing users' emotional demands. Subsequently, a spatial cross-reconstruction module is designed within the generative adversarial network to obtain more refined features. Simultaneously, a semantic correlation attention module is designed for mapping emotional needs to product images. This extracts attribute and word encodings as semantic representations to guide image generation, enhancing semantic consistency between emotional needs and visual content. Experimental results demonstrate that the proposed method achieves 92.71% accuracy in emotional need recognition and an FID of 11.88 for product images, outperforming state-of-the-art methods and delivering outstanding performance in innovative product design tasks. Inderscience Publishers - linking academia, business and industry through research

Title: Deep learning-based innovative product design driven by social network data

Authors: Changhui Wang

Addresses: Art Design Education, Zhengzhou Institute of Technology, Zhengzhou 450000, China

Abstract: Social network data has become a vital resource driving product innovation and design. Current research struggles to fully uncover users' emotional needs toward products when dealing with unstructured, high-dimensional social data, resulting in subpar product quality. To address this, this paper first employs a multi-scale attention network to analyse product emotional needs, capturing users' emotional demands. Subsequently, a spatial cross-reconstruction module is designed within the generative adversarial network to obtain more refined features. Simultaneously, a semantic correlation attention module is designed for mapping emotional needs to product images. This extracts attribute and word encodings as semantic representations to guide image generation, enhancing semantic consistency between emotional needs and visual content. Experimental results demonstrate that the proposed method achieves 92.71% accuracy in emotional need recognition and an FID of 11.88 for product images, outperforming state-of-the-art methods and delivering outstanding performance in innovative product design tasks.

Keywords: innovative product design; social network; deep learning; emotional needs analysis; generative adversarial network; GAN.

DOI: 10.1504/IJICT.2026.153619

International Journal of Information and Communication Technology, 2026 Vol.27 No.49, pp.59 - 78

Received: 02 Jan 2026
Accepted: 06 Feb 2026
Published online: 18 May 2026 *

Title: Deep learning-based innovative product design driven by social network data

Keep up-to-date