Handwritten Signature Generation Using Denoising Diffusion Probabilistic Models with Auxiliary Classification Processes

Automatic signature verification has been widely studied for authentication purposes in real life, but limited data availability still poses a significant challenge. To address this issue, we propose a method with a denoising diffusion probabilistic model (DDPM) to generate artificial signatures tha...

Full description