SDAKD: 面向超分辨率生成对抗网络的学生判别器辅助知识蒸馏
SDAKD: Student Discriminator Assisted Knowledge Distillation for Super-Resolution Generative Adversarial Networks
AI总结：
针对超分辨率GAN部署难、知识蒸馏中学生生成器与教师判别器容量不匹配的问题，提出引入学生判别器、采用三阶段训练加特征图蒸馏的SDAKD方法，在GCFSR和Real-ESRGAN上均优于现有方法。
摘要：
生成对抗网络（GANs）在图像超分辨率等生成任务中表现优异，但其计算需求较高，难以部署在资源受限设备上。知识蒸馏是GAN压缩领域颇具前景的研究方向，但由于学生生成器与教师判别器之间存在容量不匹配问题，有效训练更小的学生生成器颇具挑战。本文提出学生判别器辅助知识蒸馏方法SDAKD，这是一种全新的GAN蒸馏方案，通过引入学生判别器缓解上述容量不匹配问题。SDAKD采用三阶段训练策略，并在后两个训练阶段融合了适配后的特征图蒸馏方法。我们在两款性能优异的超分辨率GAN——GCFSR和Real-ESRGAN上对SDAKD进行了评估。实验结果表明，该方法相较基线方法与当前最优（SOTA）的GAN知识蒸馏方法均取得了稳定的性能提升。SDAKD的源代码将在论文录用后公开。
英文摘要：
Generative Adversarial Networks (GANs) achieve excellent performance in generative tasks, such as image super-resolution, but their computational requirements make difficult their deployment on resource-constrained devices. While knowledge distillation is a promising research direction for GAN compression, effectively training a smaller student generator is challenging due to the capacity mismatch between the student generator and the teacher discriminator. In this work, we propose Student Discriminator Assisted Knowledge Distillation (SDAKD), a novel GAN distillation methodology that introduces a student discriminator to mitigate this capacity mismatch. SDAKD follows a three-stage training strategy, and integrates an adapted feature map distillation approach in its last two training stages. We evaluated SDAKD on two well-performing super-resolution GANs, GCFSR and Real-ESRGAN. Our experiments demonstrate consistent improvements over the baselines and SOTA GAN knowledge distillation methods. The SDAKD source code will be made openly available upon acceptance of the paper.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
