Global context with discrete diffusion in vector quantised modelling for image generation

• المؤلفون الرئيسيون: Hu, Minghui, Wang, Yujie, Cham, Tat-Jen, Yang, Jianfei, Suganthan, Ponnuthurai Nagaratnam
• مؤلفون آخرون: School of Computer Science and Engineering
• التنسيق: Conference or Workshop Item
• اللغة: English
• منشور في: 2023
• الموضوعات: Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision; Visualization; Image Resolution
• الوصول للمادة أونلاين: https://hdl.handle.net/10356/172658

The integration of Vector Quantised Variational AutoEncoder (VQ-VAE) with autoregressive models as generation part has yielded high-quality results on image generation. However, the autoregressive models will strictly follow the progressive scanning order during the sampling phase. This leads the ex...