Diffusion-based Aesthetic QR Code Generation via Scanning-Robust Perceptual Guidance

• DOI: 10.48550/arxiv.2403.15878
• Publication Date: 2024-03-23
• Authors: Jia-Wei Liao; Winston Wang; Tzu-Sian Wang; Li-Xuan Peng; Cheng‐Fu Chou; Jun-Cheng Chen
• Keywords: Code (set theory)

QR codes, prevalent in daily applications, lack visual appeal due to their conventional black-and-white design. Integrating aesthetics while maintaining scannability poses a challenge. In this paper, we introduce novel diffusion-model-based aesthetic code generation pipeline, utilizing pre-trained ControlNet and guided iterative refinement via classifier guidance (SRG) based on the proposed Scanning-Robust Loss (SRL) tailored with mechanisms, which ensures both scannability. To further improve preserving aesthetics, propose two-stage pipeline Perceptual Guidance (SRPG). Moreover, can enhance of generated by post-processing it through Projected Gradient Descent (SRPGD) technique SRL proven convergence. With extensive quantitative, qualitative, subjective experiments, results demonstrate that approach generate diverse codes flexibility detail. addition, our pipelines outperforming existing models terms Scanning Success Rate (SSR) 86.67% (+40%) comparable scores. The combined SRPGD achieves 96.67% (+50%). Our will be available https://github.com/jwliao1209/DiffQRCode.