FedGAI: Federated Style Learning with Cloud-Edge Collaboration for Generative AI in Fashion Design

TL;DR

FedGAI tackles the challenge of enabling collaborative fashion-design sketch generation while protecting designers' IP. It proposes a three-pronged solution: a lightweight edge GAN for sketch generation, GAN compression for on-device acceleration, and federated learning to fuse multiple designers' styles without sharing data. Key innovations include AdaIN_SD for style alignment, MRFM for multi-resolution feature fusion, and FedDecorr regularization to handle data heterogeneity, all integrated within a privacy-preserving cloud-edge workflow. Empirical results on a designer-collected dataset show FedGAI achieves multi-styled sketches with quality on par with human designs and demonstrates robust, scalable style fusion across varying client counts, with significant gains in efficiency over hand-drawn workflows.

Abstract

Collaboration can amalgamate diverse ideas, styles, and visual elements, fostering creativity and innovation among different designers. In collaborative design, sketches play a pivotal role as a means of expressing design creativity. However, designers often tend to not openly share these meticulously crafted sketches. This phenomenon of data island in the design area hinders its digital transformation under the third wave of AI. In this paper, we introduce a Federated Generative Artificial Intelligence Clothing system, namely FedGAI, employing federated learning to aid in sketch design. FedGAI is committed to establishing an ecosystem wherein designers can exchange sketch styles among themselves. Through FedGAI, designers can generate sketches that incorporate various designers' styles from their peers, drawing inspiration from collaboration without the need for data disclosure or upload. Extensive performance evaluations indicate that our FedGAI system can produce multi-styled sketches of comparable quality to human-designed ones while significantly enhancing efficiency compared to hand-drawn sketches.

Figures (12)

• Figure 1: An overview of our FL-based style fusion system FedGAI.
• Figure 2: Comparison of different methods in model parameters and training data size.
• Figure 3: The lightweight GAN deployed on each client. Dashed lines indicate the flow of gradients for training the sketch generator.
• Figure 4: An overview of the MRFM module.
• Figure 5: An overview of our GAN compression method.